Summary

May 12, 2025 News Release

US: Major Companies Violate Gig Workers’ Rights

The future of work is already here. Workers around the world are increasingly hired, compensated, disciplined, and fired by algorithms that can be opaque, error-prone, and discriminatory; their faces, office badge swipes, email exchanges, browsing histories, keystrokes, driving patterns, and rest times are scanned to monitor performance and productivity; even activities outside working hours, such as health and fitness habits, social media usage, and attempts to organize are susceptible to employer surveillance and scrutiny.

Many of these practices have been normalized, if not pioneered, by digital labor platforms, which recruit workers to perform jobs or “gigs” offered through their apps or websites. In 2021, the International Labor Organization counted over 777 active digital labor platforms facilitating web-based microtasks, taxi, and delivery services globally, spanning from India and South Africa to the United States. Within ten years, the number of platform companies has multiplied nearly six times, from the 142 platforms recorded in 2010. Many of these platforms–489 of the 777–are dedicated to providing ride-hailing and delivery services. But their business model is rapidly spreading across multiple sectors of the global economy, including hospitality, health care, and software engineering. The majority of platform companies (79 percent) are situated in G20 countries. Within the G20 countries, 37 percent of platform companies are based in the US and 22 percent are in the European Union.

The US is home to one of the largest markets for labor allocated through digital platforms. In 2021, the Pew Research Center, a nonpartisan think tank, found that 16 percent of people in the US have worked for a digital labor platform at least once. Thirty-one percent of current or recent workers reported that this was their main source of income. Food delivery is the most common form of digital labor, followed by performing household tasks, providing rides, grocery delivery, and package delivery.

The growth of digital labor platforms, fueled by the promise that workers are free to set their own schedules and be their own boss, has undermined decades of US labor law regulation and enforcement, denying workers hard-won rights to an adequate standard of living and safe and healthy working conditions.

The business model of major platforms in the US exploits the prospect of flexible work to classify platform workers as independent contractors. This sleight of hand enables platforms to avoid certain taxes and employer obligations, saving labor costs. Under US federal and most state labor laws, independent contractors are not entitled to wage and labor protections guaranteed to employees, such as minimum wage, overtime pay, unemployment insurance, workers’ compensation, and paid sick leave. These protections are critical to ensuring the right to an adequate standard of living, and safe and healthy working conditions.

Despite forgoing these protections, many platform workers do not have the same freedoms that independent contractors are supposed to have, such as the freedom to negotiate basic aspects of their work, including how much they are paid. Instead, the platforms studied in this report–Uber, Lyft, DoorDash, Instacart, Shipt, Favor, and Amazon Flex–unilaterally set pay rates and deny avenues for wage negotiations. All except Amazon Flex use opaque and ever-changing algorithms that keep workers in the dark about how their pay is calculated.

Independent contractors should also be able to meaningfully control their opportunities for profit and loss. In reality, their earnings are frequently at the mercy of complex algorithmic systems that regulate the frequency and profitability of job requests offered to each worker based on their compliance with performance metrics, such as maintaining high customer satisfaction and job acceptance ratings or making on-time deliveries.

Without labor protections or bargaining power, platform workers, particularly those who work full time, are vulnerable to wages below living and minimum wage standards, wage theft, income insecurity, physical injury while on the job, and unexplained firings without meaningful recourse.

Human Rights Watch interviewed 95 platform workers in the US between 2021 and 2023, in the states of Alabama, California, Florida, Illinois, Massachusetts, Michigan, North Carolina, Ohio, Texas, Wisconsin, Oregon, Washington, and New York. Human Rights Watch also conducted a survey in 2023 of 127 platform workers in Texas, which has some of the country’s weakest labor protections.

The nonprobability survey, the first of its kind in the state, is not generalizable to other workers, but provides systematically collected case-level data on subjects that are typically obscure. It found that the median wage among those surveyed was just US$5.12 per hour after deducting work-related expenses and nonwage benefits that form part of employee remuneration. This is nearly 30 percent below the federal minimum wage, which has not kept pace with productivity or inflation since 2009, and roughly 70 percent below the living wage required for single adults with no dependents to meet their rights in the state.

Click to expand Image

Ninety-five of 127 survey respondents said they struggled to afford housing in the last year, while the majority said they struggled to afford food, groceries, electricity, and water. Forty-four respondents said that they would not be able to cover a $400 emergency expense right now, while another 64 respondents said that they would take on credit card debt or borrow from family or friends to cover the expense.

Click to expand Image

Alejandro G., a rideshare driver in Houston, told Human Rights Watch that he was making “really good money” when he first started out a few years before.[1] But “everything started getting tight” in June 2021, as rides became scarce, wait times between rides increased, and his share of passenger fares declined.[2] “The pay is not stable. There are hours where I make $20 per hour, and there are hours where I make $2 per hour,” Alejandro said.[3] He fell behind on his rent, car payments, and cell phone bill, and sold his computer and other personal belongings to a pawn shop to make ends meet. He is diabetic but cannot afford health insurance. He explained his daily struggle with constant financial insecurity:

There are nights where I am getting home at 4 or 5 a.m. in the morning, and I don’t like to work those hours, but I’m trying to scrounge up another $50 or $60. I’m just drained, emotionally drained…. All you want to do is go home, but if you go home, you won’t make any more money, but then if you stay out on the road, you just get more tired and it’s just a constant cycle.[4]

Financial distress is not the only job hazard platform workers face; they also navigate threats to their safety, health, and wellbeing on the road and in their vehicles. Rideshare drivers and delivery workers face risks of accidents and physical injuries related to traveling long hours and distances. Over one in three survey respondents said they had experienced at least one work-related car accident, while about a quarter said they had suffered from work-related injuries.

Click to expand Image

Working alone can also increase the risk of exposure to carjackings and other violent crime: Uber, for example, documented 24,000 physical assaults and threatened assaults against its drivers from 2017 to 2020. Two platform workers that Human Rights Watch interviewed had been threatened at gunpoint, while another was carjacked. Several workers reported harassment and discrimination: among the 127 workers Human Rights Watch surveyed, 54 said that they had experienced verbal abuse while working, 34 said that they experienced racial discrimination, and 16 said they had experienced sexual harassment or assault.

Click to expand Image

Significant gaps in labor and social protections in the US exacerbate these risks. Platform workers classified as independent contractors are generally ineligible for workers’ compensation in the event of work-related injuries, financial support in the event they lose their job, paid medical leave in the event of illness, and paid family leave in the event they have to take time off to care for their loved ones. Of the 127 workers Human Rights Watch surveyed in Texas, 48 said that they did not have health insurance, while 21 said they were enrolled in Medicaid, the government-funded health insurance program for people on low incomes. None of the companies pay the requisite taxes to support these workers’ use of the health care system as they would have to if they were treated as employees.

In February 2021, Debra W. tripped over a curb and fractured her arm while delivering groceries for Shipt in College Station, Texas.[5] Her injury forced her out of work for two-and-a-half months. Shipt helped with some of her medical expenses and paid for some of the hours she would have worked if she was not injured. But it was not enough to cover her bills. “I got majorly behind,” she told Human Rights Watch.[6] She said she was relying on a “wing and a prayer” to avoid repossession of her car, and to keep a roof over her head.[7] “You get to that point of, okay, what do you do? Do I pay for my car and continue to work? Or do I pay for my house and not pay for my car, and not be able to work?”[8]  

In October 2020, Julia B. was carjacked while on a package delivery route for Amazon Flex.[9] As she was closing the back door on the passenger side of the car, her car suddenly took off; somebody had followed her and waited for an opportunity to jump into the driver’s seat. The police recovered her car that evening, but the packages she was supposed to deliver that day, her keys, wallet, and cell phone were gone. The car was also scratched and dented on both sides. Julia B. took a few days off to recover from the incident, but these were unpaid and she had no access to medical support after the trauma of a carjacking.[10]

Platform workers also face the threat of being fired or “deactivated” by an algorithm; in effect being temporarily or permanently denied the ability to make a living on the platform. Digital labor platforms exercise virtually unfettered discretion over who they deactivate from the platform, for what reasons, and for how long. Of the 127 workers Human Rights Watch surveyed in Texas, 65 said that they were “fearful” or “very fearful” of being deactivated, and 40 said they had experienced deactivation at least once. Of those that have been deactivated, nearly half said that the platform ultimately cleared them of wrongdoing, indicating a high error rate.

Click to expand Image

The automated process lacks meaningful due process, even though workers stand to lose a critical means of livelihood. They are generally required to appeal any deactivation-related disputes through the platform’s app or via email with call center workers that have little authority to overturn the original decision. Most of the workers Human Rights Watch surveyed who were deactivated said they were either not given an explanation of the decision or given an inadequate explanation. Almost all said they struggled to understand the process for appealing deactivations.

Platform workers’ status as independent contractors also denies them their rights to organize and collectively bargain with their employers for better wages, benefits, and working conditions. Federal legislation that would establish a legally recognized path for platform workers to form a union–known as the Protect the Right to Organize (PRO) Act–has stalled in Congress after failing to advance in the US Senate. Workers have organized ad-hoc campaigns against missing tips, arbitrary deactivations, and shoddy safety practices, but this labor-intensive work is largely volunteer and difficult to sustain without the institutional backing of established unions and the financial and human support that provides.

The poverty, financial insecurity, and lack of economic mobility experienced by many platform workers contrast sharply with the enormous share of capital that leading digital labor platforms have captured. Uber, the leading provider of rideshare services in the US with 76 percent market share, recorded a revenue of $43.9 billion in 2024, a 17.96 percent increase from the previous year. The US market alone accounted for over 62 percent of this revenue. Uber recorded a net income of $9.8 billion for the year, and described the fourth quarter of 2024 as its “strongest quarter ever.” As of April 2025, Uber’s market capitalization stands at $169.41 billion.

DoorDash, with 67 percent share of the food delivery market, recorded a 24 percent year-over-year revenue growth in 2024, recording $10.72 billion in revenue.[11] As of April 2025, it is valued at $81.03 billion, a 51.83 percent increase since 2024.[12] The ballooning of these companies’ revenues even as their workers struggle is made possible by a regulatory framework that permits companies to not only exploit the labor of platform workers but also reduce their tax obligations, leaving public coffers with fewer resources to invest in services essential to human rights, including education, health care, and social security.

Based on tax data from the Census Bureau’s Nonemployer Statistics, Human Rights Watch estimates that Texas could have collected over $111 million in unemployment insurance contributions between 2020 and 2022 from platform companies if rideshare, delivery, and in-home platform workers had been classified as employees. The actual shortfall is likely much higher when accounting for unreported income.

Key Recommendations

Strengthening employment classification standards is a critical first step towards protecting the rights of platform workers. These standards, which determine which workers are eligible for key wage and labor protections, should be consistent with the ILO’s recommendation against “disguised employment relationships,” where an employer "treats an individual as other than an employee in a manner that hides his or her true legal status as an employee.”

One such standard is the “ABC test,” which presumes a worker is an employee unless the employer shows that the worker is: a) free from its control and direction; b) performing work outside the usual course of its business; and c) independently established in that trade, occupation, or business. The US Congress should pass the PRO Act, a federal bill that would extend the right to form a union to workers classified as employees under the ABC test. To ensure that workers benefit from just and favorable working conditions and are paid a living wage sufficient to provide an adequate standard of living, covering the costs of everyday expenses on food, water, housing, energy, health care, childcare, clothing, sanitation, transportation, education, and other expenses relevant to ensuring a decent living in the local context, Congress should also update the federal minimum wage and regularly adjust it to account for inflation.

The US should also develop regulation that protects platform workers from abusive forms of surveillance and algorithmic management. A pair of Senate bills–the No Robot Bosses Act, which establishes safeguards against automated employment decisions, and the Exploitative Workplace Surveillance and Technologies Task Force Act, which establishes an inter-agency taskforce on employer surveillance and workplace technologies–are a step in the right direction, and Congress should pass them. But there is a need for more ambitious legislation that pairs comprehensive privacy protections with safeguards against algorithmic management. The US should establish a federal data protection law that sets meaningful limits on collecting and using the personal data of workers. It should also ban the use of algorithmic management practices that pose an unacceptable risk to workers’ rights, including the use of dynamic pricing-style algorithms to determine wages, ratings-based work allocation systems, and fully automated deactivations. For practices that incur manageable risks, the US should mandate minimum transparency requirements, human rights impact assessments, and effective remedies for workers harmed by these systems.

Related Content

May 12, 2025 News Release

US: Major Companies Violate Gig Workers’ Rights

Regulation Needed to Guarantee Living Wages and Benefits, Decent Working Conditions

Glossary

Algorithm: A sequence of instructions that tells a computer how to perform certain tasks.

Algorithmic Impact Assessment: A tool for forecasting, surfacing and assessing the potential impacts of an algorithmic system on a community or society, usually conducted before a system is deployed.

Artificial intelligence (AI): There is no agreed definition of AI, but it is generally applied as an umbrella term to refer to a wide range of data-driven technologies used to perform seemingly complex tasks, including via machine learning and logic-based methods.

Dasher: People working in a delivery driver capacity for US online food delivery company DoorDash.

Employee: A person who works for an employer and is entitled to benefits and protections under US employment laws. 

Favor Runner: Texan food delivery company Favor refers to its delivery drivers as ‘Runners’.

Gamification: Design techniques that prompt platform or app users to engage with the technology in a particular way, for instance through incentive schemes, rewards systems, or ‘nudges’ that incentivize particular behaviors.

Gross hourly wage: Hourly wage after work-related expenses and employer contributions to social security and other nonwage benefits, and before employee income tax deductions or social security contributions. 

Independent Contractor / Self-Employed: A person who provides services maintaining control over the method of work while the payer only oversees the outcome, impacting both tax/social security obligations and employment status. 

Living Wage: A wage sufficient to provide an adequate standard of living, covering the costs of everyday expenses on food, water, housing, energy, health care, childcare, clothing, sanitation, transportation, education, and other expenses relevant to ensuring a decent living in the local context. 

Misclassified worker: An individual who is incorrectly categorized as an independent contractor instead of an employee, often to reduce labor costs and avoid providing benefits, circumventing internationally agreed labor law. 

Remuneration: The total amount of money that a worker receives from an app, including tips, generally per gig. This amount is equivalent to the total cost of an employee to an employer.

Remuneration after expenses: Work-related earnings after subtracting work-related expenses, also referred to as the costs of production.

Shipt Shopper: Personal shoppers who shop for and deliver groceries to customers within a set timeframe via the Shopper app from US platform company Shipt.

Social insurance: Social security schemes financed through direct contributions, typically from both employers and workers. Also referred to as contributory programs.

Social assistance: Social security schemes financed through general taxation or the general budget. Also referred to as non-contributory programs.

Social security: As set forth in international human rights law, a range of programs, whether funded from contributions or through general taxation, that encompass at least nine areas of support: health care, sickness, older age, unemployment, employment injury, family and child support, maternity, disability, and survivors and orphans.

Social security contributions: Compulsory payments paid to the government or a private entity that confer entitlement to receive a (contingent) future social benefit.

Taxes: Required payments of money to governments that are used to provide public goods and services for the benefit of the community as a whole. 

Transportation Network Company (TNC): A business that connects passengers with drivers of private vehicles through digital platforms or apps. TNCs, such as Uber and Lyft, facilitate on-demand transportation services, allowing individuals to book rides with independent drivers rather than traditional taxis or public transit.

Recommendations

To the United States Department of Labor (DOL)

• Clarify that, under Department of Labor rules establishing the economic realities test for assessing employment classification, all platform workers who are currently misclassified are eligible for protection under the Fair Labor Standards Act (FLSA).

• Meaningfully enforce the FLSA to ensure that all platform workers receive the minimum wage and wages for all hours worked. Workers should be paid on time and for all work completed.

• Ensure that all digital labor platforms comply with FLSA recordkeeping requirements, including maintaining records of workers’ active and waiting time, average and median hourly pay rates, and all updates to wage calculation formulas.

• Expand pay ratio disclosure requirements to cover private companies and platform workers within a given company. An expanded rule should require both public and private companies to disclose the pay ratio between the highest and median wages of platform workers providing services for a given company. By increasing transparency around income distribution across employment types, this disclosure would reveal wage disparities and offer platform workers valuable information to strengthen their bargaining positions.

To the US Congress

• Pass the PRO Act, which amends the National Labor Relations Act (NLRA) to strengthen collective bargaining protections and broadens its coverage to include workers that are classified as employees under the ABC test.
• Develop and pass a comprehensive data protection law that contains meaningful protections for platform workers, including restrictions on collecting workers’ personal data and using such data for work allocation and performance management functions.
• Develop and pass legislation that protects workers from abusive algorithmic management practices. Such legislation should ban practices that pose an unacceptable risk to workers’ rights, including dynamic wage-setting algorithms, ratings-based work allocation, and fully automated account deactivations. For practices that pose manageable risks, such legislation should mandate transparency and notice requirements, human rights impact assessments, and effective remedies when these practices cause harm.  
• Review and increase the federal minimum wage (currently $7.25 per hour) to a living wage sufficient to provide an adequate standard of living, covering the costs of everyday expenses on food, water, housing, energy, health care, childcare, clothing, sanitation, transportation, education, and other expenses relevant to ensuring a decent living in the local context.
• Develop and pass legislation to strengthen and adapt social security systems for non-standard forms of employment such as platform work. Ensure that platform workers are covered by unemployment and workers' compensation insurance and have access to affordable quality health care.
• Ensure that platforms are responsible for paying social security contributions on behalf of all workers and, where state benefits are not available, encourage them to provide benefits, such as medical and disability insurance, in line with local best practice.
• Overhaul the unemployment insurance program to protect all workers during job loss.
• Pass the Family and Medical Insurance Leave (FAMILY) Act, which would provide all workers up to 12 weeks of partial income when they take time for their own health or to care for children and support family members.
• Require platform companies to publish data on the earnings of platform workers on an annual basis, broken down by order and time spent working (including waiting time). Platform companies' ownership of these data, and their lack of transparency about it, creates information asymmetry between the platform companies and all other stakeholders.

To the Occupational Safety and Health Administration (OSHA)

• Regularly collect and publish data on workplace injuries for platform workers, including disaggregated by race and gender.

• Investigate workplace safety in the platform economy and extend existing health and safety legislation to include platform workers.

• Based on this research, require platform companies to adopt clear health and safety safeguards for their platform workers.

To the Federal Trade Commission (FTC)

• Investigate platform companies operating in the US and undertake enforcement actions to ensure platform workers can access and enjoy the human rights and labor rights they are entitled to, under international human rights law and standards.

• Make publicly available the findings of any investigations into the platform economy in full, including information that may be used in enforcement actions.

• Require risk forecasting and mitigation from platform companies using automated or algorithmic processes in critical functions to manage their workforce, e.g., in managing and allocating jobs, determining pay, suspending or deactivating accounts. This may include human rights impact assessments, algorithmic impact assessments, and similar tools.

To US State Governments

• Reform current unemployment insurance criteria, using the “ABC” test to determine eligibility.
• Raise state minimum wages to at least a living wage sufficient to provide for an adequate standard of living.
• Take additional steps to ensure that all workers can readily access unemployment benefits when needed, and that benefit levels appropriately reflect the need to sustain an adequate standard of living.
• Develop and pass legislation to protect the rights of platform workers, including a ban on the use of dynamic-pricing algorithms to set earnings, mandatory disclosure of work-related automation to workers, and effective remedies for deactivation.
• State legislators should clarify that platform companies are obligated to provide workers’ compensation to all their workers, either through laws that cover employees as well as independent contractors who work for them, or by classifying workers as employees.

To the State of Texas

• Rescind the 2019 Marketplace Contractor Rule and require digital labor platforms to pay state unemployment insurance taxes for their workers.
• Repeal the Transportation Network Companies Law, and permit localities to regulate digital labor platforms.
• Raise the state minimum wage to at least a living wage sufficient to provide for an adequate standard of living.

To Digital Labor Platforms

• Use clear and fixed formulas for calculating earnings, such as fixed time and distance rates or commission-based pay based on a fixed percentage of the order total.
• Establish clear and effective processes for workers to request a record of all data collected and generated about them for the purposes of task allocation, pay calculations, and performance monitoring and evaluation, as well as all decisions and other action taken as a result of such data.
• Adopt policies to protect workers from risks arising from their work and take proactive measures to protect and promote workers’ health and safety.
• Ensure that platform workers are paid a living wage.
• In determining changes in pay structure, consult extensively with workers and worker organizations, who can provide insight into the effort it takes to complete certain types of orders.
• Establish clear and effective processes for deactivation. Platforms should provide workers with a comprehensive written statement outlining the reasons for deactivation, in an accessible format that workers may access even if their account is suspended or terminated, alongside an explanation of the process through which the worker can appeal the deactivation. Platforms should address any such appeals promptly, and provide adequate compensation for lost wages in the event that the deactivation was in error or inadequately substantiated.

To the International Labor Organization (ILO)

• Collaborate with member states to establish international standards for platform work, encompassing fair wages, social security, and labor rights. Facilitate dialogue between governments, employers, and workers to develop policies that address the unique challenges faced by platform workers.

• Initiate capacity-building initiatives to support platform workers to advocate for their rights.

Methodology

This report is based on research Human Rights Watch undertook in the United States between May 2020 and September 2023, and subsequent analysis of government data and platform work under international human rights law and standards.

Data on platform workers are scarce. This scarcity stems from the classification of most platform workers as independent contractors, thereby exempting companies from reporting requirements for traditional employees regarding the number of workers engaged and their pay.

The Bureau of Labor Statistics (BLS) tries to capture information on the platform workforce through the Contingent Worker Supplement to the Current Population Survey (CWS), which measures workers in alternative work arrangements such as independent contracting, on-call arrangements, and employment arrangements through temporary agencies or contracted firms. However, the survey reflects only the type of work individuals do as their main or sole job and does not capture supplemental work. To complement this information, research centers, worker groups, and academic institutions have collected data through surveys at both state and national levels. Most state surveys have focused on California and East Coast states. To date, and to our knowledge, no survey focusing specifically on Texas has been conducted.

To address this gap, Human Rights Watch conducted a survey of 127 workers in Texas and interviewed 65 workers within the state. The survey was not intended to capture a representative sample of platform workers in Texas, and its findings are indicative but cannot be generalized to all platform workers in the state. Nor can it be used to generalize about experiences of platform workers across the US, as pay rates, cost of living, and minimum wage protections vary by state and within states. However, the survey results largely echo accounts from people Human Rights Watch interviewed, as well as results from larger national and state-level surveys.

To help evaluate how platforms collect workers’ data and use technology, Human Rights Watch additionally interviewed 30 US platform workers outside Texas. Of those, 15 interviews were with platform workers in California; the other 15 were with workers in Alabama, Florida, Illinois, Massachusetts, Michigan, North Carolina, Ohio, Wisconsin, Oregon, Washington, and New York.

This testimony is instructive because data collection and algorithmic management practices on platforms, such as pay structures, ratings, and rewards schemes, are largely similar across the US.

Interviews

Human Rights Watch staff conducted in-person research in Austin, Dallas, Houston and San Antonio, Texas, between May and July 2021. One interview was conducted over email. All other interviews were conducted by telephone or a video conferencing platform. Two interviews were conducted in Spanish by a researcher fluent in that language, and all other interviews were conducted in English.

All workers interviewed provided verbal informed consent to participate. Workers who participated in in-person interviews received a $20 gift card as partial reimbursement for travel and related expenses they incurred to participate in the interview. Interviews ranged from 45 minutes to 2 hours and were conducted while workers were not actively signed on to any of the platforms they work for.

To protect workers’ identities, especially as many expressed fear of retaliation from their current or prospective employers, Human Rights Watch uses pseudonyms for all workers interviewed in this report, using a first name and an initialized last name (e.g., Leslie M.). All individuals who appear in this report with full last names are identified by their real names and affirmed that they were willing to have their names and the contents of their interviews published.

Since there are few places where platform workers congregate or organize, it is difficult to identify groups of workers. For that reason, many workers Human Rights Watch interviewed in Texas as well as other states were identified through Rideshare Drivers United, an independent non-profit association of rideshare drivers, and the Gig Workers Collective, which was at the time of research a non-profit group comprising current and former platform workers.

Others responded to a flyer soliciting interviews with platform workers that Human Rights Watch and Gig Workers Collective posted on Craigslist, Facebook, and Reddit. Human Rights Watch also used snowball sampling (where interviewees introduced Human Rights Watch to their colleagues). Frequently, Human Rights Watch was not aware of the identity of the worker or their work history prior to the interview. The interviews reflected in this report capture the experiences of workers working for a range of platforms across different states and metro areas.

For background and contextual information, Human Rights Watch interviewed relevant subject matter experts, including organizers, restaurant owners, lawyers, union representatives, journalists, and academic researchers. The report also draws extensively on publicly available secondary information to corroborate findings gathered through interviews, including reports from nongovernmental organizations, government and academic studies, legal proceedings and rulings, books, and relevant local and national reporting.

The federal and state governments do not keep comprehensive data on platform workers. Human Rights Watch relied on a range of sources, including surveys conducted by labor organizations, universities, and private corporations. Human Rights Watch followed methodologies developed by The New School University and the UC Berkeley Labor Center to analyze public data to estimate the number of platform workers in Texas, the share of workers who rely on public assistance programs like SNAP or Medicaid, and the loss in Texas state revenue in the form of forgone taxes by digital labor platforms who do not pay unemployment insurance taxes when classifying platform workers as independent contractors.

Survey

To complement the interviews, Human Rights Watch conducted a survey among a sample of 127 platform workers in Texas, primarily based in the metropolitan areas of Austin, Dallas, and Houston.

In preparation for the survey, we sought guidance from researchers at the University of California Los Angeles (UCLA), Cornell University, and Columbia University to learn about methodologies they had employed in surveys with platform workers in California, New York City, and across the country. We focused on survey design, implementation strategies, and whether workers were compensated for their time. The consultations revealed significant challenges in reaching platform workers, citing workers’ reluctance in participating due to fear of retaliation and lack of trust in individuals outside their communities. Some surveys relied exclusively on online distribution tools (e.g., Facebook), while others utilized hybrid approaches with both online and in-person outreach. All reviewed surveys provided compensation or an opportunity for compensation to participants (e.g., through a raffle). The table beow summarizes the surveys and compensation methodologies used.

Overview of Consultations on Similar Surveys

Informed by consultations and the challenges identified in reaching potential survey respondents, we opted for a survey methodology centered on participatory worker outreach. This approach entails hiring platform workers to actively promote and encourage other workers to participate in the survey.

To implement this methodology, Human Rights Watch collaborated with Rideshare Drivers United (RDU), an independent non-profit association of rideshare drivers based in California with members across the country.

RDU was primarily responsible for recruiting survey participants by hiring one local platform worker in each of the three cities to recruit survey participants via social media and in person. The hired platform workers helped spread the word by sharing flyers about the survey with a QR code in restaurants where workers pick up food, in supermarkets where workers shop, or places where rideshare drivers wait for customers (e.g., an airport). All surveys were completed through a web browser (Typeform). Cornell University used this approach successfully in New York City in partnership with the Workers Justice Project, a New York-based worker group that mobilized platform workers in their communities.

There were two criteria for taking the survey: the respondent had to be over the age of 18 and had to have worked for at least one month for a food or grocery delivery or rideshare app in Texas. Human Rights Watch excluded responses if the respondent did not affirm that they were over 18 and freely participating in the survey, if the respondent reported that they had never worked for an app or worked less than one month for an app, if the respondent failed certain tests for reliability embedded into the questionnaire, if more than one response was associated with a worker’s contact information, or if the geographic information obtained did not match zip codes in Texas.

The resulting sampling methodology was a combination of purposive, convenience, and snowball sampling methods. The purposive aspect ensured that respondents were platform workers meeting the aforementioned criteria. Membership in RDU was not required; any platform worker in Texas involved in rideshare or grocery delivery apps could participate. Convenience played a role in recruitment, with organizers seeking respondents at various restaurants or online, recruiting merely based on the condition of the respondents being a platform worker and fulfilling the two criteria. The snowball element involved workers encouraging peers to partake in the survey. While the resulting sample is not representative, it is indicative of the working conditions of respondents who work for apps and serves to validate the information obtained from the interviews.

Data collection took place between December 2022 and March 2023 using the Typeform platform, with 48-item questionnaires available in both English and Spanish. A total of 1,028 users opened the survey, and 476 responded to at least one question. Cleaning steps to ensure responses met the required criteria and exclude fraudulent entries, as well as removing outliers, left us with 127 survey responses, of which 100 were completed in English and 27 in Spanish. On average, verified respondents completed the survey in about 20 minutes. We used the interquartile rule (1.5 times IQR) to remove outliers in the responses to questions asking about earnings. Each respondent provided informed consent regarding the purpose, use, and confidentiality of their survey responses. Verified survey respondents received a $15 gift card for completing the survey.

The questionnaire, developed by Human Rights Watch and RDU, consisted of 48 questions covering various aspects, including earnings, work-related expenses, work history, workplace safety, social protection, deactivation, and the impact of low earnings. The data collected through the survey, such as data on wages, could not be independently verified by Human Rights Watch.

Human Rights Watch conducted the analysis of the survey results. The survey instrument and the flyer to recruit participants are available on the Human Rights Watch website.

Digital Labor Platforms Studied

Some of the workers Human Rights Watch interviewed and surveyed engaged with apps for more than six years; others started during the Covid-19 pandemic. The eight apps studied in this report, in alphabetical order, are Amazon Flex (owned by Amazon.com, Inc.), DoorDash (DoorDash, Inc.), Favor (H-E-B Grocery Company), Instacart (Maplebear Inc.), Lyft (Lyft, Inc.), Shipt (Target Corporation), Uber (Uber Technologies, Inc.), and UberEats (Uber Technologies, Inc.).

Click to expand Image

Focus on Platform Workers’ Pay in Texas

Texas was selected as the primary research location for studying platform workers’ pay for several reasons. Texas is the second most populous US state, with a 2024 US census population of 31,290,831.[13] The state also has some of the largest cities in the country: Houston, San Antonio, Dallas, and Austin rank among the 11 most populous cities in the US.[14]

Most studies of working conditions in the digital platform economy have broad national coverage, or focus on high-cost cities in coastal states (such as New York City, Los Angeles, San Francisco, Seattle, and Washington, DC).[15] Comparatively few researchers have studied the particularities of platform work in southern states, and Human Rights Watch is not aware of any comprehensive study or survey that documents the conditions of platform workers in Texas.

Texas has some of the lowest indicators for adequate labor protections among US states. Oxfam’s Best States to Work Index, which evaluates states' efforts to protect, support, and remunerate workers, ranks Texas 46th among 52 US states and territories, placing it last in the southwest region.[16]

The state's performance suffers particularly in terms of compensation and workplace conditions. Notably, Texas maintains one of the lowest minimum wage rates, currently stagnant at the federal level of $7.25.[17] To put this into perspective, the minimum wage of $7.25 covers a mere 20.3 percent of what the Massachusetts Institute of Technology estimated in 2023 as necessary to cover the cost of living for a family of four with two children and one working adult ($35.75/hr).[18] Localities in Texas cannot set the minimum wage above the $7.25 state standard.[19] Texas also lacks provisions for mandatory paid leave.

Even though these wages are low and protections are few, platform workers cannot enjoy them since the digital labor platforms studied in this report generally classify platform workers as independent contractors. As a result, platform workers are not entitled to the minimum wage, workers’ compensation, anti-discrimination safeguards, and other labor protections. Moreover, Texas explicitly removed those protections from platform workers in 2019 when Texas’ state unemployment agency exempted platform companies from the state’s unemployment insurance regulations, effectively classifying all platform workers as independent contractors.[20]

Texas is also one of 26 states[21] with so-called ‘Right-to-Work’ laws, which are designed to undermine freedom of association, and a vestige of the Jim Crow era that permit employees to opt out of paying union dues, leading to weaker unions and lower wages.[22]

Though our research focuses on workers throughout the state, Texas is a geographically, economically, and demographically diverse state, and the experiences of the app workers we identified cannot represent all app workers.

Outreach to Companies and Government Officials

In March 2022, Human Rights Watch wrote letters to the seven platform companies listed above (Uber and Uber Eats are part of Uber Technologies), sharing preliminary findings from this research and requesting information and clarification regarding their practices. We received a written response from Amazon and Lyft and met with Amazon to discuss the questions and findings. In the letters, we requested information about workers in Texas (total number of workers engaged with the platforms and demographic information), pay and bonus schemes (median hourly pay before and after tips, work-related expenses, and the pay implications of drivers’ behavior and performance), ratings (implications on where, when, and how people work), support mechanisms, and financial assistance, insurance, and benefit schemes, including in the event of workplace injuries. No platform company provided any of the requested data.

In October 2022, Human Rights Watch sent follow-up letters to Uber and Lyft, seeking information about factors that guide their algorithm’s pay calculations. The companies declined to share such information.

In March 2025, Human Rights Watch wrote to each of the seven companies with further questions. In response, Lyft said “App-based work provides millions of Americans uniquely flexible work opportunities, leaving room for them to meet other goals, commitments, or obligations. It allows them to work around their many real and unpredictable commitments and their busy schedules in ways that traditional 9-5 jobs don’t provide.” Lyft also shared policy documents with Human Rights Watch. Amazon met with Human Rights Watch to discuss the report but did not give an on-record response. The other companies did not reply. The letters and their responses are available on the Human Rights Watch website.

In March 2025, Human Rights Watch wrote letters to government officials within the US Department of Labor, the Federal Trade Commission, and the Texas Workforce Commission, to share the findings of this report. As of the time of publication, the Department of Labor had acknowledged receipt but did not provide a comment on our findings. None of the other agencies had responded.

Background

Digital platform work is spreading across multiple sectors of the economy, from ride-hailing, food delivery, and caregiving, to research, content moderation, and data labeling that helps develop artificial intelligence systems.[23] In 2021, the International Labour Organization (ILO) identified a total of 777 digital labor platforms worldwide, a nearly sixfold jump from the 142 operating in 2010.[24] The majority of platform companies (79 percent) are situated in G20 countries. Within the G20 countries, platforms are largely concentrated in the United States (37 percent) and the European Union (22 percent).[25] Ride-hailing and delivery platforms made up the lion’s share: 63 percent, or 489 platforms, a roughly tenfold increase within the previous decade.[26]

Ride-hailing and delivery platforms have positioned themselves as a more convenient way of getting rides, ordering meals, and running errands, and their reliance on AI and other digital technologies as a more efficient means of generating supply and demand and managing labor costs.[27] The ILO has observed that these platforms generally forgo “traditional capital assets, such as cars, hotels or warehouses” in favor of investments in technology:

[P]latforms tend to invest instead in digital infrastructure and are overwhelmingly dependent on data, skills, ideas and physical assets provided by their users (both clients and workers). For example, Uber does not heavily invest in cars, but [as of 2021] it has been able to expand and scale in 69 countries at an unprecedented pace (within 11 years of its creation). It has 26,900 employees and 5 million drivers, who either own or lease cars, with the majority of them being labelled as self-employed or “driver-partners”. Uber orchestrates its services through its app, which is its “linchpin” (algorithmic management), by matching customers with drivers: its key assets are the network of users (drivers and consumers), data and the brand.[28]

This business model requires vast computational resources and expansive access to worker and consumer data to manage and control fares, routes, compensation, and schedules.[29] But many of the business and labor practices this model automates are hardly new. Just-in-time scheduling–matching workers to available work on demand–was developed by Japanese carmaker Toyota in the early 1970s to minimize manufacturing delays.[30] Paying platform workers per gig harks back to an early 20th century labor practice that targeted immigrant women, who were paid per piece rather than by the hour to knit for garment manufacturers, and earned half what women factory workers made.[31] Nonstandard employment–a hallmark of the digital platform economy–has been rising globally for decades.[32]

While platforms tout their commitment to innovation, their extensive use of technology has failed to staunch familiar concerns about the impact of their underlying business on workers, consumers, and the environment. Like the taxi industry, rideshare platforms have attracted criticism for increasing congestion, pollution, and risks to public safety. A 2020 study by the Union of Concerned Scientists found that taking a non-pooled ride-hailing trip generates about 47 percent greater emissions than a private car trip.[33] A 2021 study by researchers at Carnegie Mellon University found that non-pooled ride-hailing trips were more polluting than private car trips and responsible for a greater increase in road congestion, traffic noise, and the likelihood of car accidents.[34] Researchers partially attribute these impacts to a practice known as “deadheading”: the time that drivers spend on the road without passengers in the vehicle (for example, while traveling to the next pickup or looking for the next ride).

The ways that platforms deploy technology, however, has shielded many of their practices from scrutiny and accountability. How rideshare platforms profile consumers to algorithmically price fares, for example, is largely unknown since both the code and the data used to train their algorithms are considered proprietary. But researchers at George Washington University managed to reverse engineer ride-hailing fares in Chicago based on anonymized data that the city had compelled platforms to disclose. Their 2021 study of the data found that Chicago neighborhoods “with larger non-white populations, higher poverty levels, younger residents, and high education levels are significantly associated with higher fare prices.”[35] Although these findings indicate consumer discrimination, it is impossible to replicate or scale this analysis without access to similar data in other locations. Human Rights Watch has also not come across any study of whether dynamic pricing in the food and grocery delivery sectors has discriminatory effects.

This report examines a key technological development that has marked the rise of rideshare and delivery platforms, and broad swaths of the digital platform economy: the use of AI and other data-driven technologies to manage and control workers. It examines how algorithmic management practices deployed by these platforms are implicated in labor rights abuses, while simultaneously providing cover for these abuses. Peeling back the layers of technological complexity and obfuscation, this report also reveals how the labor practices of major platforms are fueling the interconnected crises of informal work and economic inequality.

Digital Platform Work in the United States

The United States is one of the largest markets for digital labor. In 2021, a national and representative Pew Research Center survey found that 16 percent of US adults have worked for a digital labor platform at least once.[36] Hispanic adults are more likely than any other racial or ethnic group to have worked for these platforms in the US, and young adults between 18 and 29 are more likely to have done so than older adults.[37]

Ride-hailing and delivery services are the most popular and visible forms of platform work: Seven percent of Pew respondents have worked for food delivery platforms, five percent for ride hailing platforms, and four percent for grocery delivery platforms.[38] But platform work is gaining a foothold in other industries: US hospitals and healthcare facilities, for example, are increasingly reliant on app-based nursing platforms to ease staffing shortfalls,[39] despite reports that they have degraded working conditions for nurses, and exposed patients to accidents, medication errors, and neglect.[40]

In recent years, several platform companies have experienced significant market share and revenue growth. Uber captured 76 percent of rideshare sales in the US in early 2024. Its market capitalization stands at $169.41 billion as of April 2025, and 156 million people used Uber or Uber Eats in 2024, a 13.8 percent increase from the previous year.[41] In the company’s fourth quarter 2024 results release, CEO Dara Khosrowshahi stated: “Uber ended 2024 with our strongest quarter ever […]. We enter 2025 with clear momentum and will continue to be relentless against our long-term strategy.”[42] Similarly, CFO Prashanth Mahendra-Rajah affirmed: “Record demand in both Mobility and Delivery helped us grow Gross Bookings faster than the high end of our guidance, and we closed out 2024 exceeding our three-year outlook […]. We believe we remain undervalued despite these strong fundamentals, and plan to be active and opportunistic buyers of our stock.”[43]

Lyft, the second largest rideshare provider in the US, accounted for 24 percent of all rides.[44] As of April 2025, Lyft’s market capitalization reached $5.21 billion.[45] In 2024, Lyft recorded $5.8 billion in revenue, up 31 percent from the previous year.[46] Lyft reported closing the fourth quarter of 2024 with “record Gross Bookings, significant margin expansion” and its “first full year of GAAP profitability, and record cash flow generation.”[47]

In the realm of food delivery, DoorDash dominates, with 67 percent of meal deliveries in the US as of March 2024, followed by Uber Eats with 23 percent and Grubhub with 8 percent.[48] As of April 2025, DoorDash had a market capitalization of $81.03 billion, and reached $10.72 billion in revenue. In 2024, it reported $123 million in net profit.[49]

People most vulnerable to income insecurity are more likely to sign up for platform work. The Pew survey found that 25 percent of people with lower incomes have worked for platforms, compared to 13 percent of those with middle incomes and 9 percent with upper incomes.[50] They are also more likely to rely on platform work as their main source of income: 42 percent of workers with lower incomes told Pew that this was their main job over the last 12 months, compared to 31 percent of all workers.[51]

Human Rights Watch’s survey in Texas focused on rideshare drivers and food, grocery, and package delivery workers. It did not solicit responses from other types of platform workers. Hispanic adults made up the largest share of Human Rights Watch’s survey respondents (39 percent), about equivalent to the state’s Hispanic or Latino population.[52] Our respondents were generally older than Pew’s: more than half were between 30 and 49, followed by workers between 50 and 64, and workers between 18 and 29. Eighty-six workers who identified as men took the survey; the remaining 41 identified as women.

Responses to Human Rights Watch’s survey were consistent with the Pew survey, in that the large majority of its respondents are on lower incomes: among the 127 survey respondents, 52 said they struggled to cover their housing costs almost every month within the last year, while another 42 said they struggled for one to three months. About a third struggled with covering their food costs almost every month, while nearly another third struggled for one to three months.

Many platform workers are drawn to their jobs because they believe it offers them more flexibility than other types of work. According to Pew’s survey, 49 percent of current or recent platform workers said that “being able to control their own schedule” was a major reason for taking on platform work, while 35 percent said that “wanting to be their own boss” was a major reason.[53] (Human Rights Watch’s survey did not ask respondents about their reasons for choosing platform work.) The marketing pitches of major ride hailing and delivery platforms drive home the narrative of flexible work: on these platforms, they claim, workers get to set their own work hours, be their own boss, and make a living on their terms.[54]

Employment Classification

Platforms invoke this narrative to classify platform workers as independent contractors rather than employees, under federal and state labor law.[55] Under these laws, independent contractors are not entitled to the minimum wage, overtime pay, unemployment insurance, and anti-discrimination safeguards.[56] Some paid hours associated with employee status, including standby time, lunchtime, sick leave, and vacation, are unpaid hours for independent contractors.[57] Workers shoulder the costs of being injured on the job, and all work-related expenses.

While employment classification standards are in flux and vary by jurisdiction, they largely turn on the extent to which employers exercise control over their workers. Labor law experts have identified three main tests of employment classification in the US. The common law standard, which establishes a 13-factor test of whether the employer has exerted control over the “means and manner” of the worker’s work, determines whether or not a worker is classified as an employee who is entitled to collective bargaining and anti-discrimination protections under federal law, and workers’ compensation in the event of work-related injuries or illness in many states.[58] The economic realities test, which probes the degree of managerial control and entrepreneurial freedom present in the employment relationship, the skill and initiative required of the worker, and the permanency of the relationship, determines eligibility for wage and hour protections, family medical leave, and social security benefits under federal law.[59] The ABC test presumes employee status unless the worker is free from employer control and direction, performs work outside the usual course of the employer’s business, and is customarily engaged in an independent trade, occupation, profession, or business. This test, which imposes the most stringent criteria for classifying workers as independent contractors, determines eligibility for unemployment insurance protections in many states.[60] But few states apply the ABC standard other than for determining eligibility for unemployment insurance.

This complex patchwork of classification rules has made it difficult to hold employers accountable for misclassification practices that lead to labor rights abuses. In the early 1980s, many companies began to outsource work to contractors in and outside the US, in a bid to cut labor costs and improve profitability.[61] They converted secure, full-time jobs into “part-time work, temporary positions and other ‘contingent’ forms of employment” often performed by workers classified as independent contractors.[62] This form of outsourcing created downward pressure on wages and benefits “for identical kinds of work and workers,” and incentivized contractor firms to shirk labor laws. David Weil, who served as the Administrator of the Wage and Hour Division of the US Department of Labor[63] from 2014 to 2017, described observing persistent wage theft from service workers employed by subcontractors during his tenure at the DOL.[64]

Since 2014, digital labor platforms have waged a successful campaign to further weaken state classification laws, making it harder for platform workers to qualify for state-level wage and labor protections.[65] Thirty states have in the past decade passed laws that presume or designate drivers of “Transportation Network Companies (TNC)” as independent contractors.[66] Ten states have also passed laws that classify all platform workers (defined as “marketplace contractors” under the law) as independent contractors.[67]

The contested status of platform workers has left them vulnerable to exploitative wage and labor conditions. In the rideshare and delivery sectors, numerous news reports, research studies, and worker surveys have documented wages as low as $3.37 per hour;[68] missing tips;[69] draconian financial penalties for falling short of exacting performance standards;[70] arbitrary, erroneous or unexplained firings that are difficult to appeal;[71] and heightened exposure to carjackings, assaults, and accidents.[72]

Platforms collect and generate a wide range of data on compensation, account suspensions, and work-related injuries and crime. But these data are largely shielded from disclosure to US government agencies responsible for enforcing labor laws, and workers themselves. The classification of platform workers as independent contractors exempts platforms from recordkeeping and reporting obligations on wages, hours, and work-related illnesses and injuries under federal law.[73] Some platforms provide workers the option to request a copy of their personal data but not information about how their data is used to make work-related decisions, such as generating their fraud probability scores and allocating rides or orders.[74]

In the Netherlands, drivers have successfully sued Uber and Ola, another ride-hailing company, for access to such information under the EU’s General Data Protection Regulation. But the US lacks a comprehensive data protection law that would enable workers to access this data.[75]

State-Level Regulation

A handful of states have attempted to address platform-enabled employment misclassification and related labor rights abuses. In September 2019, California passed a law known as Assembly Bill 5 (“AB5”), which designates the ABC test as the default standard for determining whether a worker is an employee under the state’s Labor Code and Unemployment Insurance Code.[76] AB5 classifies app-based rideshare drivers and delivery workers as employees, entitling them to state-level wage and labor protections such as the minimum wage, overtime pay, workers’ compensation, and access to unemployment benefits.[77] Uber and its subsidiary, Postmates, as well as representatives of the trucking industry, unsuccessfully challenged AB5 in federal court.[78] In June 2024, the US Court of Appeals for the Ninth Circuit, in San Francisco, dismissed the challenge to AB5 from Uber and Postmates, stating that “the legislature perceived transportation and delivery companies as the most significant perpetrators of the problem it sought to address—worker misclassification.”[79]

Platforms also attempted to undermine AB5 by establishing a “third way” of classifying platform workers: maintaining their independent contractor status while offering curtailed labor guarantees.[80] During the November 3, 2020, state election, Uber, Lyft, and DoorDash successfully petitioned voters in California to pass a ballot measure known as Proposition 22, which exempted employers of “app-based ride-share and delivery drivers” from AB5 and related wage and employment laws.[81] Prop 22 instead established a wage calculation formula that effectively lowered the minimum wage for platform workers in the state and imposed prohibitive criteria for accessing health insurance subsidies.[82] The three companies, along with Instacart and Postmates (which was later acquired by Uber), spent over $200 million to pass the measure, the largest amount ever spent on a ballot initiative campaign in the state.[83]

In 2021, Rideshare Drivers United and PolicyLink analyzed detailed trip and earnings data on more than 12,000 rides completed by 55 rideshare drivers across the state, and found that drivers’ median take-home earnings under Prop 22 were $6.20 per hour.[84] That same year, a survey conducted by RDU and National Equity Atlas of 531 rideshare drivers found that only 10 percent of respondents were receiving Prop 22’s health insurance stipend, while 40 percent had never heard about the stipend or were unsure about whether they qualified.[85] In July 2024, the California Supreme Court upheld Proposition 22, allowing platform companies to continue misclassifying workers as independent contractors.[86]

Platforms have backed the passage of Prop 22-style laws in other states that have resisted previous efforts to weaken classification rules for platform workers. In March 2022, the state of Washington passed a law that classifies platform workers as independent contractors while establishing minimum pay guarantees, paid sick leave, and a process for them to appeal platform suspensions.[87] The law was a compromise between Uber, Lyft, and the local chapter of the Teamsters Union representing rideshare drivers in the state, although the union’s national headquarters subsequently rejected it.[88] In Massachusetts, Uber, Lyft, and others spent $17.2 million on a campaign to pass a ballot measure that was modeled after Prop 22, but the state’s Supreme Judicial Court invalidated the measure.[89]

Texas has made it virtually impossible for platform workers to qualify for employee status and related wage and labor protections. It is one of ten states that classify all digital platform workers–described as “marketplace contractors”–as independent contractors.[90] The “marketplace contractor” rule, which was established by the Texas Workforce Commission, the state’s labor agency, in 2019, exempts platforms from paying state unemployment insurance taxes for those workers.[91] The Commission reportedly adopted the rule after intense lobbying by Handy, a home repairs platform.[92]

The 2019 rule expanded a 2017 law that designates all drivers for “Transport Network Companies” as independent contractors and blocks local regulation of rideshare services.[93] The TNC law invalidated local regulation passed by city governments in Austin, Corpus Christi, Galveston, Houston, and Midland to establish licensing, safety, and background check requirements.[94]

Algorithms of Exploitation

Workers for the digital labor platforms studied in this report are assigned orders, supervised, paid, and fired by algorithms. This interconnected system of financial controls, performance metrics, and behavioral nudges is enabled by the pervasive monitoring of virtually every move that workers make, such as how they accelerate and brake on roads, how they communicate with customers, the rate at which they pick groceries, and the jobs they accept, cancel, or reject.

For the platforms, these interventions are an ongoing attempt to solve the eternal puzzle of supply and demand: matching workers to rides or deliveries in real time, while keeping customers satisfied and driving down labor costs. For workers, this form of algorithmic management harbors tremendous potential for exploitation. The algorithmic systems that govern their work are frequently opaque, making it difficult to understand how they are monitored, paid, evaluated, and fired. This lack of transparency contributes to a working environment that incentivizes long, grueling hours for low and unpredictable pay, with detrimental effects on their rights to a decent living and safe and healthy working conditions, while hampering accountability for such abuses. It also creates potential discrimination for some workers, notably those with disabilities, who may not meet a platform’s rigid automated or algorithmically determined “productivity” goals, without reasonable adjustments.[95] 

At the same time, by using opaque, proprietary algorithms to shield how they manage and control workers from public scrutiny, platforms preserve the narrative that they offer workers the flexibility to work when and where they want, and the corresponding legal status of these workers as independent contractors. This legal classification in turn prevents workers from unionizing or relying on other aspects of US labor law, such as minimum wage and workplace safety standards, to counter the abusive effects of how they are algorithmically managed and controlled.

Surveillance

Digital labor platforms require platform workers to consent to the collection and analysis of a wide variety of personal and sensitive data as a condition of employment.[96] Human Rights Watch examined the privacy and data collection policies of Uber, Lyft, DoorDash, Instacart, Shipt, Amazon Flex, and Favor, and interviewed workers to understand how these policies applied.[97]

The types of data collected are largely split across four categories: 1) data about workers’ locations and movements; 2) data about their behavior and performance; 3) communications between workers and customers, and workers and the platform; and 4) biometric data.

Location Data

The collection of geolocation data is crucial to the functionality of most ride hailing and delivery apps, and the majority of platforms compile extensive records about their workers’ locations and movements, both when they are working, and sometimes when they are not working.

Uber, Lyft, Instacart, DoorDash, and Shipt specify that they collect precise location data about workers, such as their geolocation via GPS coordinates, when they are actively engaged with the app (e.g., on screen while looking for or completing a ride or a delivery) or letting it run in the background (e.g., when they may not be working but have not switched off the app).[98] Shipt also activates location tracking on a worker’s device when they have scheduled a work shift in advance and their “planned work window is approaching,” in order to dispatch geographically “relevant offers for potential deliveries.”[99] Favor and Amazon Flex collect precise location data, but do not specify when this collection starts and stops.[100]

All seven companies say that they use geolocation data to track workers’ progress on rides or deliveries and to detect fraudulent activity. Shipt, Instacart, and Lyft specify that they use this data to investigate whether someone other than the worker is using their account. Uber says that it uses location data gleaned from drivers’ selfies to detect this type of fraud, but only in the UK.

Uber, Lyft, Instacart, DoorDash and Shipt also use this information to match workers with rides and deliveries. Uber adds that it uses this information to customize the amount it charges customers and pay workers for each ride or delivery, while Shipt relies on this information to ensure “accurate payment” for workers.[101]

Behavioral Data

Many platforms also collect and analyze vast amounts of information about how workers behave when they are on the clock.

DoorDash provides one of the more granular accounts of how the performance of workers is tracked and monitored.[102] Whenever a Dasher logs onto their app, the company records the date, time, and worker’s location. DoorDash also records how the Dasher scrolls through the app, every page they visit, and every link they click. From the moment a Dasher accepts a delivery order to the time they complete it, nearly every move is tracked: what routes they travel, when they pick up an order and drop it off, the order amount and tips received, and the overall time the delivery takes. Each time a worker declines or cancels a delivery is also tracked. DoorDash uses this information to generate or supplement performance metrics such as the Dasher’s acceptance and cancellation rates, customer ratings, and the number of deliveries they make over time.[103]

Similarly, Uber, Lyft, Instacart, and Shipt collect and analyze extensive information about when and how workers navigate their apps, the rides or deliveries they accept, cancel, and reject, and their progress throughout every ride or delivery.[104]

Amazon Flex, which requires workers to sign up for delivery shifts rather than specific deliveries, collects data that meticulously tracks their progress throughout a shift. From the time they report to their assigned warehouse, to when they pick up parcels, begin a delivery route, return from the route, and drop off undelivered parcels, Amazon workers’ behavior can be analyzed via time, date, and location. Amazon Flex also monitors whether workers are driving, walking, or running on a route, and when they are using “driver assistance technologies” (such as autopilot mode).[105]

While tracking workers in this way allows employers and customers insight into individual jobs, it creates a disproportionate risk for workers with disabilities and older workers, who may be flagged or penalized for not completing tasks within a rigid timeframe, where there are no procedures in place to request for reasonable accommodations for their disability or other needs.[106]

Additionally, platforms monitor specific aspects of workers’ behavior that are relevant to the nature of the work performed. Uber, Lyft, and Amazon Flex, for example, collect information about vehicle speeds and how drivers brake and accelerate on roads to monitor whether they are driving safely.[107] Instacart and Uber Cornershop, the company’s grocery delivery service, measure the speed at which workers pick groceries, but workers Human Rights Watch interviewed were unclear about whether or how this metric affected the dispatching of shop requests.[108] Uber has also disclosed that it analyzes both driver and rider data, such as trip history and “reported incident rates,” to “predict and help avoid pairings … that may result in increased risk of conflict.”[109]

Communications

Workers’ communications are also closely monitored. Lyft notifies drivers and riders that their calls will be recorded before the recording begins.[110] DoorDash uses an undisclosed third party to monitor and analyze chat and text messages between Dashers and customers for fraud, violations of the company’s terms of service, and quality and training purposes.[111] Shipt records text messages, phone calls, email exchanges and in-app communications between shoppers and customers, as well as shoppers and the company.[112] Instacart does not disclose whether it monitors shoppers’ communications with customers, but both its privacy policy and independent contractor agreement notify shoppers that it records their calls with the company.[113] Amazon Flex collects all communications between drivers and Amazon personnel, including phone calls to support services.[114] Amazon wrote to HRW on May 10, 2025, to clarify that “this type of data collection is used for incident investigations and has data deletion policies that apply”.

Biometric Data

All the platforms studied in this report collect personally identifiable information such as workers’ names, phone numbers, and email addresses as well as certain biometric data. Uber, Lyft, Shipt, Instacart, DoorDash, and Amazon Flex have outfitted their workers’ apps with a type of facial recognition technology known as facial verification to check whether the person signing on to the platform matches the identity of the worker on file.[115] This technology requires the collection and storage of workers’ facial images, via photo ID documents and selfies submitted by the worker through the app.

Relying on facial recognition technology to verify identity increases the risk of discrimination to workers of certain racial or ethnic identities, women, and people with disabilities, potentially putting them at risk of losing earnings or even wrongful termination.[116] In 2019, the National Institute for Science and Technology, a federal government laboratory, found that commercially-available facial verification algorithms produce more false negatives for darker skinned people and women than for lighter skinned people and men.[117] (False negatives occur when the algorithm mistakenly concludes that the person is not who they say they are.) NIST has also warned that “[w]hen image quality degrades, false negatives are expected to increase”; for example, when workers try to complete facial verification checks while driving or in poor lighting.[118] Amazon wrote to Human Rights Watch on May 10, 2025, to clarify that “[f]or Amazon Flex, delivery partners are not terminated using automated tech. There is always a human involved.”

Workers from Instacart and Uber have shared on social media anecdotal accounts that they have been temporarily shut out of their accounts or permanently deactivated after they failed the platform’s facial verification checks.[119] In the United Kingdom, the Independent Workers’ Union of Great Britain (IWGB) has documented at least 35 Uber drivers who have been deactivated from the platform apparently due to errors with the company’s facial verification technology. One of these drivers has filed an employment tribunal claim against Uber challenging the deactivation.[120]

Opacity

Although platforms harvest vast amounts of data about platform workers, they disclose very little to workers, regulators, and the public about how they use these data to set pay, track, and monitor performance, and influence working conditions.

Workers cannot set or negotiate their rates for giving a ride or shopping and delivering an order, as a true independent contractor can. Instead, the platforms increasingly dictate the terms of their compensation through opaque and ever-changing algorithms. A 2021 Pew Research survey found that more than half of the platform workers have a poor understanding of how digital labor platforms determine how they are paid.[121]

Platforms relying on black box algorithms to calculate base pay (i.e., pay before tips and applicable promotions or bonuses) include:

1. Uber, which rolled out a new pay structure in February 2022 known as “Upfront Fares,” which calculates drivers’ earnings based on an algorithmic assessment of factors such as estimated length and distance of the trip, the distance to pickup, and real-time demand at the destination.[122] This dynamic, real-time formula was a major departure from its longstanding method of calculating fares based on fixed time and distance rates and applicable fare multipliers such as surge pricing.[123] Uber declined to provide the full list of factors that guides the algorithm’s calculations.[124]

2. Lyft, which has since 2022 gradually rolled out a similar model known as “Upfront Pay” to calculate drivers’ earnings, relying on time, distance, travel to pickup, real-time demand for rides at the driver’s location, and other unspecified “market factors.”[125] Lyft declined to provide Human Rights Watch with the full list of factors that guide their algorithm’s pay calculations.[126]

3. Instacart, which calibrates the amount it pays shoppers for each “batch” (the company’s term for an order or a group of orders) based on the “size and complexity” of the orders in each batch.[127] This pay model takes into account factors such as delivery time and distance, travel to pickup, and the presence of heavy grocery items.[128] It is unclear what other factors guide the company’s assessment of “size and complexity.” When the company announced these new pay features on July 20, 2023, it assured shoppers that its algorithmic pay calculations will not fall below a minimum base pay of $4 for each batch.[129] This minimum pay figure is a sharp decline from the $7 to $10 per batch minimum established in 2019.[130] Before 2018, Instacart would pay a commission of $1 to $14 per order (depending on time and place), as well as a fee for each grocery item (usually around $0.40).[131]

4. Shipt, which ties shopper pay to its algorithm’s estimation of the time and effort it would require to shop and deliver the order. Factors influencing pay include “the size of the store, driving time, the likelihood of [order] substitutions and their complexity.”[132] It has assured shoppers that pay calculations will not fall below a minimum base pay of $16 per hour.[133] This figure, however, encompasses compensation for both the worker’s labor as well as their work expenses, such as gas and car maintenance. Shipt’s compensation formula was far more straightforward before 2021: it paid shoppers $5 for each order completed, along with a 7.5 percent commission of the total order amount.[134]  

5. DoorDash, which offers workers the ability to toggle between two payment options. The “Earn per Offer” mode calculates base pay on a per order basis, according to the “estimated time, distance and desirability of the order.”[135] The company declined to provide additional information about how it measures “desirability.” The “Earn by Time” mode, which was established in June 2023, provides workers a fixed hourly pay rate that varies according to location.[136] The pay rate only covers active time (meaning the time between accepting the order and dropping it off), and does not cover wait time between orders.[137] In “Earn by Time” mode, workers have little say over the orders they pick up, as they are permitted to decline only one offer per hour.[138]

6. Favor: The base pay for each delivery, known as a “Favor,” starts at $2.10 and “can increase for different types of Favors.” The criteria that Favor’s algorithm uses to adjust pay are unclear.[139]

Amazon Flex is the outlier among these companies. It pays Flex drivers a flat hourly rate ranging from $18 to $25 depending on location.[140] It is unclear if or how Amazon uses the data it collects on workers to set these rates.

Opaque pay algorithms make it harder for workers to assess how their compensation is measured, whether particular jobs are worth their time and effort, and the factors responsible for changes to their pay. “Large orders would pay the same as tiny orders,” said Rachel L., a Shipt shopper in Dallas. “We have no idea how the algorithm works. No one knows how they calculate the time the order takes to fill.”[141] Isabel H., a Shipt shopper, questioned whether these assessments take into account real-world conditions such as traffic or lengthy wait times at the checkout counter when stores are understaffed.[142] Jacob F., a former shopper and delivery courier, questioned whether this lack of transparency was consistent with his employment classification. “Why is this info kept in such a vault for an independent contractor? I should know. That allows you to determine whether a shop is worth your time, whether a company is worth your time.”[143]

Many workers Human Rights Watch interviewed set daily or weekly income targets but found that the amount of time and work it took to reach these targets would vary in unpredictable ways, making it difficult to budget and make ends meet. Mary A., an Instacart shopper, said she usually works 10- to 12-hour days to hit her daily income goal of $100. Some weeks, however, she averages only $70 a day despite working the same amount. “I am behind my rent because of this,” Mary said.[144] “I have struggled since October [2020] to make enough money to keep a roof over my head.”[145] Jacob F. had a similar experience. “Some days I know I need to make X amount of money to pay a bill or pay rent, and I don’t make it. I’m not getting the shops or the tips. How do you go home and sleep?”[146]

Among the 127 workers Human Rights Watch surveyed in Texas, a majority reported experiencing food insecurity in the previous year. Nearly one in three reported experiencing difficulty affording food and groceries nearly every month.

Incentivization

Some platforms use gamification techniques to prompt workers to engage with their apps and incentivize them to behave in a certain way.[147] For example, in 2017, the New York Times revealed that Uber was using game design techniques to nudge its drivers into staying on the app for longer periods of time and accepting more jobs. [148] These techniques impact a worker’s earnings directly (by providing cash incentives to an individual upon the completion of a task) or indirectly (by prioritizing or de-prioritizing some workers over others for various ride or delivery requests). The cumulative effect of gamifying the design of apps creates additional barriers to transparency about wages, meaning that workers know even less about how, where, and why they have been compensated. Gamification contributes to financial uncertainty and unpredictability, and also strongly influences when, where, and how workers work.

Bonus Schemes

Cash bonus schemes are designed to recruit new workers as well as incentivize existing ones to drive and deliver more frequently during peak periods or in areas with high demand. Bonuses can be a boon for workers, particularly when demand for their labor is high. But bonuses can also be a trap, luring workers to work more for less pay while obscuring their true earnings.

Uber, for example, entices drivers to work in particular areas or stay on the road through “Surges” (earnings multipliers for completing rides in areas with high demand) and “Quests” (cash bonuses drivers attain when they complete of a set number of rides within a specific time period).[149]

Since 2016, drivers have reported multiple incidents where they travel to surge areas only to realize that the surge has disappeared, or that it is lower than Uber had initially promised.[150] Luis P., an Uber driver that Human Rights Watch interviewed in Houston, explained:

The map with surge pricing entices drivers to travel to certain areas, but when you go there, there is no surge. The map feature predicts where and when it will be busy, and does not always reflect the reality of demand…. They are like puppet masters and they psychologically manipulate you.[151]

Chasing surges can rack up time, mileage, and wear and tear on the driver’s car, increasing their work expenses with no guarantee of higher earnings. Veena Dubal, a law professor at the University of California Irvine and labor rights expert, has observed that this incentive structure induces drivers to take bets on “work activities connected to earnings that limit choice and present high financial risk”; a practice she calls “algorithmic gamblification.”[152]

Quest is another manifestation of algorithmic gamblification. When Human Rights Watch interviewed Cindy L., a veteran rideshare driver in San Francisco, Uber was offering her a menu of Quest options, including opportunities to make an additional $150 after completing 50 trips within a four-day period, $205 after 60 trips, or $270 after 70 trips. “You have no context for how many [much work] that is if you haven’t driven for that platform. It’s just a lot more hours and effort than you think,” she said.[153]

Daniel T., a rideshare driver based in Orange County, California, explained that looming Quest deadlines frequently make him stay on the road longer, and accept rides that are not financially worthwhile:

[I] feel pressured to work more. I have to take rides no matter what, even if there is no surge, and even if the ride is not in a busy area. Sometimes the pickup location is 10, 15 or 20 minutes away from me, but I have to take the ride.[154]

The possibility of a sizable cash reward lures drivers like Cindy L. and Daniel T. to accept Quests without adequately considering their impact on their flexibility, schedule, and overall earnings. “People are addicted to the cash payment into your account. It’s really like a video game,” Cindy L. added.[155]

Drivers barely making ends meet may feel additional pressure to accept Quests. Alejandro G., a rideshare driver in Houston, told Human Rights Watch that his non-bonus earnings with Uber are so low that completing Quests is the only way he can make a decent living. “I get glimmers of hope with Uber when I get $300 per day, and then it becomes $70 per day,” said Alejandro. Without Quest, he said, working as a rideshare driver would not make financial sense.[156]

Uber is not the only platform that deploys bonus schemes to induce compliance with desired performance metrics. Lyft’s “Bonus Zones” operate like surges, and “Ride Challenges” like Quests.[157] DoorDash has a referral bonus scheme that rewards existing Dashers and people they enlist to work for the platform with a cash bonus when the new Dasher completes the required number of deliveries in a specified timeframe.[158]

Mae D., a platform worker in Portland, Oregon, started working for DoorDash in May 2021 when she learnt from her daughter that the company was offering a $600 sign-on bonus to new Dashers after completing 250 deliveries in two months.[159] “To get the bonus, I had to take a lot of orders and deliveries on which I lost money,” Mae said.

She blamed her “own lack of knowledge” for taking delivery requests that were financially unsound; for example, a request that sent her to Scappoose, Oregon, a 60-mile journey roundtrip for which she was paid $10. But she was also torn between her impulse to reject a money-losing proposition and the growing realization that “I absolutely won’t be able to get the bonus if I was picky about my orders.”[160]

Human Rights Watch analyzed Mae D.’s earnings records (308 DoorDash orders over 39 days) during the bonus period and found that, with the $600 bonus included, her average hourly pay was $14.04/hr, barely the minimum wage in Portland, which was $14/hr in 2021, and below the $18.72/hr that MIT considered to be a living wage for a single worker with no dependents in Portland.[161] Human Rights Watch calculated the hourly pay rate by deducting the Internal Revenue Service (IRS) mileage rate for each mile driven for work. This rate, set by the US federal government, estimates expenses associated with using a car for business purposes. It includes expenses like fuel, maintenance, insurance, and depreciation. The rate is adjusted every year; the 2021 mileage rate was $0.56 per mile driven.[162]

Without the bonus, her hourly rate would have been $8.50/hr, well below Portland’s minimum wage. Human Rights Watch also found that her hourly pay exceeded the local minimum wage only on 13 of the 33 days she worked, even when accounting for the bonus pay.

Click to expand Image

Dispatching

The algorithms that some platforms rely on to match workers with rides and orders also heavily influence earnings, since they dictate the number of ride or order requests dispatched to each worker, the time and distance traveled to fulfill each request, pay rates, and wait time between requests.

These algorithms function as middle managers that control the flow and type of rides or deliveries offered to each worker based on their performance and behavior. Through the review of documents and worker interviews, Human Rights Watch identified three factors that influence dispatching: 1) how frequently workers accept rides or deliveries and the types of work they accept; 2) how highly customers rate them; and 3) elaborately structured rewards programs that unlock individualized modifications to the dispatching algorithm upon compliance with desired performance metrics.

Ride or Delivery Acceptance Patterns

Some platforms take into account whether workers are likely to accept a ride or delivery request based on their past behavior.[163] Uber’s Privacy Notice states that it matches drivers to riders based on factors “such as likelihood to accept a trip based on their past behavior or preferences.”[164] DoorDash explains that it tries to “anticipate the types of offers that are more likely to be accepted and present them to the most relevant Dasher,” indicating that it does not only analyze overall acceptance rates but possibly also factors such as the pay rates or mileage each worker is willing to accept.[165] It is also experimenting with a rewards program that conditions priority access to higher-paying orders based on the Dasher’s acceptance, completion, and customer ratings, which is discussed below.[166]

Researchers at Columbia University have raised concern that this type of predictive matching may facilitate individualized pay discrimination: namely, the risk “that it learns each worker’s reservation wage, the lowest rate each is likely to accept, and then tailors offers to each accordingly.”[167] Tyler A., an Uber driver in Cleveland, Ohio, explained to Human Rights Watch how this might happen, particularly as Uber rolls out its “Upfront Fares” model:

I think it’s entirely possible that drivers’ earnings lose any standardization and become entirely individualized. With drivers now being able to see where a ride is going and how much it will pay, the companies will have tons of new and specific data on how much a driver is willing to accept, how far they’ll drive, where they’ll drive … and can use that to offer the bare minimum they predict the driver will accept, or at least offer different rates to different drivers and create a race to the bottom.[168]

Customer Ratings

All but one of the platforms studied in this report prompt their customers to rate their driver or delivery worker after a ride or a delivery. Customer ratings are highly subjective, and can be misleading and even discriminatory. However, platforms rely extensively on this metric to assess, reward, and penalize workers’ performance. These ratings can spell the difference between a decent living and sub-minimum wages, and keeping and losing their jobs. Although their right to work hangs in the balance, there are very limited ways for workers to appeal ratings they deem incorrect or unfair.

Of all the platforms studied, Shipt’s ratings system has the bluntest and most direct impact on their workers’ ability to make a decent living. Shipt assigns shoppers an overall customer rating calculated on a five-point scale, and based on an average of the most recent fifty ratings they receive from customers.[169] The platform’s algorithm offers the most lucrative orders to the highest rated shoppers, leaving behind fewer and lower paying orders for everyone else. A shopper’s on-time percentage, calculated based on their last fifty orders, also “influences order offers.”[170]

Shoppers told Human Rights Watch that less-than-perfect ratings can lead to an order drought and a steep fall in earnings.[171] In September 2023, the orders offered to Elizabeth S., a shopper in Kalamazoo, Michigan, slowed to a trickle after she received a pair of four-out-of-five star ratings. “Starting yesterday [September 10, 2023], my offers are a lot less. It was one of the worst Sundays I ever had. Typically, I make $200 and I didn’t even make $150. And today, it’s usually $250 and I’m just scraping to get to $150,” Elizabeth S. said.[172] She appealed to Shipt to reverse one of the four-star ratings but the appeal was denied.

In April 2021, a customer penalized Isabel H., a Houston shopper, with a one-star rating because the grocery store did not have many of the items they requested.[173] “They rated that against me even though it was not my fault.”[174] The following week, “I didn’t get many offers because of my ratings drop,” Isabel said. “I am on the schedule for 9 a.m. to 1 p.m., but I would only be able to do 2 or 3 orders.”[175] This was because the app offered her orders “nobody else wanted,” and sent her fewer orders overall. Her decline in earnings caused her to fall behind on her car payments and utility bills.[176]

Shoppers Human Rights Watch interviewed said that, under this system, their ability to make a living is at the mercy of the whims of customers. Jacob F., a former Dallas shopper, recalled receiving a four-star rating from a customer that did not respond to his questions about whether he could make replacements for grocery items that were out of stock. “I really don’t know why I got rated a four. I did not do anything different with this customer that I did with any other customer…. I imagine they weren’t happy with the substitutions, but it is hard to function when the customer does not respond.”[177]

It takes considerable time and effort for shoppers to recover financially from a low rating. Isabel H. said that it took her two weeks to recover from her one-star rating. Jacob F. recalled that it took him a “solid month” to accumulate enough five-star rated orders before Shipt dropped a low rating from his overall rating calculation.[178] “That takes a lot of work,” Jacob F. said.[179]

Shipt says that it automatically drops a shopper’s lowest individual rating from the calculation of their average, but shoppers told Human Rights Watch that this still leaves very little margin for error.[180] Shoppers can also fill out an online form to appeal a low rating, but it is unclear how extensively the company investigates each appeal.[181] In November 2022, Shipt said that it was experimenting with a new feature that would automatically forgive low ratings from customers that routinely fail to provide reasons for their rating.[182]

Prior to May 2022, Instacart maintained a similarly punitive rating system that offered batches first to the highest rated shoppers, followed by the next highest.[183] Four Instacart shoppers told Human Rights Watch that this system effectively imposed harsh financial penalties for a handful of low ratings, which are highly subjective and sometimes awarded for reasons beyond the shopper’s control.[184] Separate investigations by the Los Angeles Times and ABC News found that as few as two low ratings caused shoppers to lose hundreds of dollars in potential earnings.[185]

In May 2022, Instacart relaxed its ratings requirements, announcing that shoppers need only maintain a rating of 4.7 or above to enjoy priority access to batch requests. However, shoppers with lower than a 4.7 would continue to “see a batch a few minutes after shoppers who have a 4.7 rating or above see it.”[186] In July 2022, Instacart revised its dispatching policy again to award highest priority to shoppers that not only maintain a minimum 4.7 rating but also complete the greatest number of deliveries. This development, announced as part of a rollout of its shopper rewards program Cart Star, is discussed below.[187]

Uber, Lyft, and DoorDash require workers to maintain a minimum customer rating in order to remain active on their platforms.[188] Their rewards programs also impose minimum ratings requirements in exchange for in-app features and incentives that offer the possibility of higher earnings.

Rewards Programs

Some platforms have rolled out rewards programs that offer individualized modifications to dispatching rules in exchange for meeting various performance targets. This system is akin to frequent flier programs by airlines: workers unlock reward tiers once they meet that tier’s targets. But failing to meet these targets leads to the denial of incentives that can have a negative effect on earnings and reduce scheduling flexibility. This stick-and-carrot approach is designed to incentivize workers to complete as many highly rated jobs for the company as possible, particularly during peak periods. 

Instacart and DoorDash, for example, assign dispatching priority to workers in their highest reward tiers. Under Instacart’s Cart Star program, shoppers with “Diamond Cart” status, the highest reward tier, enjoy priority access to batches at nearby stores, and exclusive access to batches before stores open. To earn Diamond Cart status, shoppers must complete 200 orders every three months, and maintain a 4.7 customer rating throughout this period.[194]

DoorDash is piloting a program in multiple US locations that dispatches the highest-paying deliveries to Dashers based on their rewards tier. To attain Platinum status and enjoy “top priority,” Dashers must maintain a 70 percent acceptance rating, a 95 percent order completion rating, a customer rating of 4.7, and have completed a hundred deliveries in the last 30 days. Platinum Dashers are also prioritized for large catering orders over all other Dashers, which can lead to “bigger tips and higher earning potential.”[195]

The decision to comply with these performance targets is fraught with financial risk. Both programs disadvantage workers who are selective about the orders they accept, or who toggle between multiple apps in search of the most lucrative orders, since they are less likely to complete the required number of deliveries and at higher risk of losing access to the highest paying orders. On the other hand, workers who strive to preserve dispatching priority are under pressure to accept as many orders as possible, even if some of those pay below the living wage or even the minimum wage.

Uber and Lyft take a different approach, rewarding their highest rated drivers with some degree of control over where they pick up and drop off rides. Uber drivers with “Diamond” and “Platinum” status are entitled to an in-app feature known as “Area Preferences,” which enables drivers to start and finish rides in their area of choice for two hours a day.[196] According to the Rideshare Guy, a rideshare and delivery industry blog, the feature provides “more flexibility” to eligible drivers while reducing deadhead time (“no more trips to the middle of nowhere and wondering how to get back”) and wear and tear on the driver’s vehicle.[197]

Uber’s drivers have extremely limited insight into rider demand at any given time or location, and certainly not enough data to allow them to make an informed choice about whether or not they opt to use the Area Preferences functionality. Activating this feature in areas with less demand can lock drivers into long wait times between rides, and fewer bonus earnings opportunities.

Lyft drivers with “Gold” and “Platinum” status are entitled to additional uses of an in-app feature known as “Destination Mode,” which is supposed to match drivers with rides closer to their destination of choice.[198] Human Rights Watch has previously found that this feature can be counterintuitive, matching drivers with rides in the opposite direction of where they want to be, as long as the app determines that they have enough time to get there by their designated time preference.[199] Both Uber’s Area Preferences and Lyft’s Destination Mode belong to a long line of gamification techniques that ultimately reinforce the platform’s control over where, when and how their employees work.[200]

Deactivation

Workers have expressed grave concern about the ability of platforms to suspend them temporarily or indefinitely based on a wide range of infractions and without adequate transparency or due process; a largely automated process known as “deactivation.” For example, half of the platform workers Human Rights Watch surveyed were “fearful” or “very fearful” of being deactivated and only about 1 in 10 stated they were not fearful at all. Over a third of respondents said they had been deactivated at least once.

Human Rights Watch’s analysis of the deactivation policies of Uber, Lyft, Shipt, Instacart, DoorDash, and Amazon Flex reveals that workers may be deactivated for a broad range of reasons.[201] Theft, abuse, harassment, discrimination, and violence are grounds for deactivation.[202] Performance-related issues–such as late or incomplete deliveries,[203] failing to follow a customer’s delivery instructions,[204] cancelling too many jobs,[205] low customer ratings,[206] and poor vehicle maintenance[207]–may also lead to temporary suspension, which can be reversed upon the completion of course requirements and other corrective measures. Failure to comply with these requirements can lead to permanent deactivation.

Workers frequently struggle to understand why they are deactivated, or how to appeal their deactivation. Among the dozens of Human Rights Watch survey respondents that had been deactivated at least once, a third were not given an explanation for their deactivation, while a majority said that the explanation offered was unclear or “not very” clear. Over a third said they were unable to figure out how to appeal the deactivation, while a majority said that the process was unclear or “not very” clear.

Nearly half of the respondents who had previously been deactivated said that the company ultimately found no wrongdoing, indicating that their deactivation was in error. Lucy M., whom Human Rights Watch had interviewed prior to conducting the Texas survey, said that Favor deactivated her account in November 2020, alleging that she had failed to complete a delivery. Lucy scrambled to submit pictures and location data to Favor’s support team showing that she had completed the delivery. “I was so stressed out. Favor was my only source of income,” she said.[208] Favor reinstated her account a week later, but the damage was done. “Missing one week can completely throw off the whole month,” Lucy said. “I fell behind on rent.”[209]

Even workers capable of demonstrating compliance with the platform’s performance requirements are vulnerable to deactivation for reasons beyond their control. On November 23, 2020, Sandra C., a Shipt worker in Austin, Texas, received an email notifying her that more than 10 percent of her deliveries were late, and that she needed to take an online course on completing deliveries on time. Screenshots Human Rights Watch reviewed showed that she had only been late once in over 60 deliveries that month. But Sandra figured it was easier to take the course than to appeal, and completed the course on November 25.[210]

On November 30, 2020, Shipt deactivated her account, claiming she had not completed the course. She had to retake the course twice more before Shipt reinstated her account, on December 6, 2020. The ordeal left Sandra feeling “angry and unheard.” She estimates that she lost hundreds of dollars in earnings that week. “I just barely scraped by,” Sandra said.[211]

Denied Fair Wages

All workers have the internationally recognized right to benefit from just and favorable working conditions.[212] However, most platform workers classified as independent contractors are deprived of the legal protections that ensure this right. This lack of protection has resulted in many workers enduring unstable earnings that fall far short of a living wage.[213]

Human Rights Watch considers a living wage as the minimum necessary for workers and their families to enjoy their economic, social and cultural rights, including their right to an adequate standard of living. Governments should set and regularly review legal minimum wages so that all workers, regardless of their employment contract, earn a wage that enables them to pay for goods and services essential to realizing their rights. These include everyday household expenses such as food, water, housing, energy, health care, childcare, clothing, sanitation, transportation, education, and other expenses relevant to ensuring a decent living in the local context.[214]

While there is no clear standard for what a living wage would be in the United States and in each of its states, there are various tools that universities, think tanks, and NGOs have developed for estimating national and local living wages.[215] Perhaps the most comprehensive and commonly used, the Living Wage Calculator developed by the Massachusetts Institute of Technology, estimates that in 2022, the national living wage for a family of four (two working adults and two children) was $25.02 per hour after taxes.[216] This is an average, with calculations also made for the 50 states and Washington, DC; more than 3,100 counties; and 384 metropolitan areas. The living wage calculations take nine cost categories into account: food, childcare, health care, housing, transportation, civic engagement, internet, some miscellaneous items, and income and payroll taxes. Therefore, the living wage reflects the amount a worker should receive after taxes and social security contributions have been deducted.[217]

In Texas, in 2022, when Human Rights Watch conducted many of the interviews for this report, the MIT calculator estimated the living wage to be $16.75 per hour for a single person and $23.62 per hour for a family of four, slightly lower than the national average. By 2025, these estimates had risen to $21.82 and $24.63, respectively.[218] For comparison, the Economic Policy Institute’s 2024 estimates for Texas are higher, with a $26.58 per hour for a single person and $26.46 per hour for a family of four with two children and two working adults.[219]

In this section, we examine platform workers’ wages by comparing them to legal minimum wages and the living wage calculations based on MIT’s Living Wage Calculator for Texas and the United States more broadly. While this does not imply endorsement of the living wage estimates, it serves to illustrate how current wages compare against estimates of what is required for a decent living.

To make valid comparisons between platform workers and traditional employees, we account for the costs of production, which include expenses related to the use and upkeep of vehicles, phones, and other materials essential for performing the job, as well as social security contributions typically covered by employers for their employees. Since platform workers are classified as self-employed, their remuneration is neither a net or gross wage, but equivalent to the total cost of an employee to an employer.

While platform workers’ remuneration is often compared to employees’ wages, these comparisons are misleading. Employee wages are the amount workers receive after the employer has already covered various costs, including work-related equipment and obligatory and voluntary benefits. In contrast, platform workers’ remuneration reflects the amount before any deductions, leaving them responsible for covering these expenses themselves.

To provide an accurate assessment, we start with the amount a platform worker receives from an app, including customer tips, which we refer to as the total remuneration. Tips can play an important part in workers’ remuneration. In a study conducted in Seattle by the labor group Working Washington, tips made up about a third of the pay of delivery and car sharing platform workers.[220] A 2022 study by the New York City Department of Consumer and Worker Protection found that about half of workers’ hourly pay stems from tips.[221]

As the next step in our assessment, we account for waiting time in our calculation of working hours to determine an hourly remuneration. We then deduct:

• a) Costs of production (e.g., fuel, car maintenance, phone costs), and

• b) The employer’s share of nonwage benefits, including mandatory social security contributions (commonly referred to as payroll taxes in the US), as well as voluntary social security and benefit contributions, as required by international human rights standards.

This calculation yields the equivalent of a worker’s gross hourly wage, that is before deducting income taxes and allows for comparison with federal and state minimum wage levels.

A living wage is more closely aligned with net pay, also known as take-home or pocket pay: the amount an employee receives after all employer and employee-side deductions, social security contributions, income tax, health insurance premiums, and other deductions. Because some of these deductions, like income tax or health insurance premiums, depend on various factors beyond gross hourly pay and we lack specific information for each worker, we refrain from estimating the net hourly pay. However, since living wages are intended to represent the amount after all deductions, it is evident that if the gross amount falls below the living wage threshold, it is inadequate.

We start with a review of existing research across the United States and then analyze responses from the survey Human Rights Watch conducted with 127 workers in Texas. This analysis illustrates the difference between the remuneration workers receive from an app per gig and their hourly wage after expenses and deductions.

Click to expand Image

From Seattle to New York City: Prior Research Finds Wages Below a Living Wage

Research conducted by local authorities, universities, and think tanks in Seattle, New York City, Denver, Chicago, and nationwide consistently indicates that platform workers experience high rates of economic insecurity. Generally, they earn low wages, often falling below or just above the applicable minimum wage, and significantly below a living wage.

The table on the following page summarizes several studies analyzing platform workers’ hourly remuneration and wages in various localities. These studies employ different sample sizes and methodologies to estimate the costs of production and calculate wages. Some rely on data from local governments, others use data provided by platform companies, and some conduct their own surveys. While all studies account for the cost of operating vehicles for platform work, some also consider the social security contributions that employers have to pay for employees, and that independent contractors have to cover from their remuneration.

The purpose of this compilation is not to compare hourly wage rates across platforms and jurisdictions, but to take stock of existing studies and compare them to Human Rights Watch’s findings.

Across jurisdictions, platform sectors, and companies studied - using diverse methodologies including survey data, company data, and city government data–none of the studies found earnings that meet or exceed the living wage benchmarks defined by MIT. Only a few studies found earnings at or above the respective minimum wage standards.
 

A 2020 US-wide survey by the Shift Project, a collaboration between the Harvard Kennedy School and the Economic Policy Institute, compared the wages of platform workers with employees in the service sector.[231] While both groups often grappled with low wages and precarious working conditions, the study found that platform workers earned consistently less, even after accounting for tips.[232] Approximately one in seven platform workers earned an hourly wage below the federal minimum wage, whereas no employees in the service sector reported wages below the federal minimum. Moreover, 29 percent of platform workers earned less than the state minimum wage had they been classified as employees.[233] This comparative study shows that without the labor protections granted to employees, workers are far more likely to earn low and substandard wages.

Full-time platform workers are particularly vulnerable to low and unpredictable wages. According to a Pew survey of platform work in 2021, 31 percent of all platform workers in the US, and 42 percent of those living on low incomes, rely on app-based work for their livelihoods.[234] A seven-year study of the gig economy by Boston University economist and sociologist Juliet B. Schor and a team of academics found that such workers struggle to make platform work “a viable way of living”: they are frequently below the poverty line, “lose autonomy and scheduling flexibility,” and live in fear of deactivation, bad ratings, and arbitrary pay changes.[235]

Surveys that have sampled a large proportion of full-time platform workers confirm these findings. In New York City, a survey by the city’s Department of Consumer and Worker Protection (DCWP) found that 77 percent of delivery workers using e-bikes and 56 percent of platform workers using cars performed deliveries as their main or only job, with earnings after work-related expenses at $11.12 per hour, including tips. Another study by the DCWP, based on compliance data submitted by Uber Eats, DoorDash, and Grubhub, which together make up 95 percent of the NYC food delivery market, found that, without tips, workers earned just $5.39 per hour. Both estimates are far below the state’s $15 minimum wage, even before accounting for any tax or social security contributions.[236]

In April 2024, the New York City Mayor and the DCWP Commissioner announced that, effective immediately, the city's “minimum pay rate” (referring to the hourly app pay) for app-based restaurant delivery workers would be raised to $19.56 per hour (before tips and before tax or expense deductions) and would be adjusted annually for inflation. The “pay rate” reflects the citywide $15 per hour minimum wage for most employees, plus “adjustments for differences in tax treatment between employees and independent contractors.”[237]

The new pay rule comes after years of legal debates between the city government and platform companies. In June 2023, the city government had first announced the minimum pay rule, following a months-long rulemaking process that included two public hearings and thousands of public comments. In early July, several delivery apps sued the city, seeking to stop the minimum pay rate from taking effect. In September, a state court ruled in the city's favor, allowing enforcement of a minimum pay rate to begin.[238] The apps appealed the ruling, and in late November, the Appellate Division denied the appeals, paving the way for DCWP to begin enforcing the minimum pay rate.[239]

In July 2024, Massachusetts’ Attorney General’s office reached a settlement with Uber and Lyft, mandating that starting August 15, 2024, workers must be paid at least $32.50 per hour while en route to pick up passengers or actively transporting riders. This rate is subject to annual inflation adjustments. Additionally, the companies are required to offer portable health funds, allow drivers to accrue up to 40 hours of paid sick leave per year, and secure occupational accident insurance for their drivers.[240]

Studies from other countries have found similarly low pay among platform workers. For instance, an ILO study in Chile found that platform delivery workers earned nearly 900 pesos (about $1.24) less per hour, compared to traditional delivery workers who are employees. The main reason for the difference in wages was the disparity in work hours. For the same or less pay, app workers worked about 62 hours per week, compared to the 45 hours worked by traditional delivery workers.[241] In the United Kingdom, a survey of over 500 platform workers by the University of Bristol found that half of respondents reported wages below the minimum standards for employees.[242]

The following sections outline three steps to derive hourly wages from the total remuneration apps provide. Using data from the survey Human Rights Watch conducted in Texas, we illustrate these calculations. This analysis shows that figures from companies, such as Uber’s estimate that workers earn $18.67 per hour in Houston,[243] are misleading, as they fail to account for the reality that workers must cover the costs of work-related equipment and self-employment expenses such as social security contributions. They also fail to consider the full amount of time workers spend engaged with gigs when calculating “work hours.”

Click to expand Image

Step 1 – Determining Hours Worked, Compensating Wait Time

The first step in estimating the hourly wages of platform workers is assessing the hours worked. While this is usually straightforward for employees whose working time starts when they clock in or appear at the workplace, platform workers are generally only compensated for the time they are actively delivering an order or transporting a passenger; referred to by many platform companies as “engaged” or “active” time. This approach can neglect the time spent driving back from a client’s house to the store or restaurant, or the time waiting between orders, and leaves many of their working hours uncompensated. This also contributes to long workdays and significant variations in earnings, depending on the demand for their services and the supply of workers at a given moment. Therefore, determining what constitutes work hours and compensating workers for wait time is key to arriving at a fair hourly wage that can be compared to a living wage.

For on-call workers, the US Department of Labor evaluates whether wait time is compensable on a case-by-case basis, depending on the extent to which workers have relinquished autonomy and control over the time at issue. Key considerations in this assessment include: the duration available for responding to a call, the frequency of such calls, and the latitude to decline a specific request without facing detrimental employment repercussions. The Department has emphasized that “[i]f the calls are so frequent or the on-call time conditions so restrictive that the employee cannot effectively use the on-call time for his or her own purposes, the on-call waiting time would constitute hours worked.”[250]

To date, platforms do not compensate workers for the time they spend waiting between rides or deliveries. A 2019 survey conducted in major US metropolitan areas and funded by ride-hailing platforms found that drivers, on average, spend only 54 to 62 percent of their work time with passengers in the car.[251] In Minnesota, a 2024 study suggests that this figure is 58 percent.[252] In Seattle, it is estimated that drivers have passengers in their vehicles 49 percent of their work time on the road.[253] In New York, this figure stands at 58 percent for both Uber and Lyft drivers.[254] In Chicago, drivers have passengers in the car for just 55 percent of their working hours.[255]

Platform-backed minimum wage floors in California and Washington only provide compensation for engaged time, not waiting time.[256] If platform workers were classified as employees, however, they would be entitled to compensation for waiting time under the Fair Labor Standards Act (FLSA). In Armour & Co. v. Wantock, et al., the US Supreme Court found that workers should be compensated for all of the time they are on the clock, including time spent waiting for work. The Court explained:

Of course, an employer, if he chooses, may hire a man to do nothing, or to do nothing but wait for something to happen. Refraining from other activity often is a factor of instant readiness to serve, and idleness plays a part in all employments in a stand-by capacity. Readiness to serve may be hired, quite as much as service itself…. We think the Labor Standards Act does not exclude as working time periods contracted for and spent on duty in the circumstances disclosed here merely because the nature of the duty left time hanging heavy on the employees' hands and because the employer and employee cooperated in trying to make the confinement and idleness incident to it more tolerable.[257]

In the wage analysis of the 127 respondents to the Human Rights Watch survey in Texas, we consider the time workers spend waiting as work hours. The respondents reported spending an average of 40 percent of their work hours waiting for orders or rides. A quarter of the workers said they spend more than half their work hours waiting. When accounting for waiting time, workers in the survey sample received $16.90 per hour in total remuneration from the apps.

Step 2 – Deducting the Costs of Production

The second step in estimating hourly wages is to deduct the costs of production, which include expenses related to the use and upkeep of vehicles, phones, and other essential materials. As independent contractors, platform workers are responsible for selecting, acquiring, and maintaining the equipment and services needed to connect to the apps and perform deliveries or rides. They are responsible for all costs associated with their vehicles, including fuel, insurance, and depreciation, which may vary widely. They also shoulder repair costs when vehicles require maintenance, and platform companies do not reimburse these costs.

Surveys and data collected by local governments and researchers show that such expenses can consume a significant share of workers’ remuneration. A 2017 study of nationwide administrative tax records by staff of the US Department of the Treasury found that 70 percent of platform workers’ remuneration was spent on work-related expenses, leaving them with average after-expense earnings of just $5,700 per year before taxes and social security.[264]

A 2020 study with a representative survey commissioned by San Francisco’s Local Agency Formation Commission, a state regulatory body responsible for overseeing the changes to the boundaries of the city and special districts, and conducted by the University of California at Santa Cruz, found that work expenses can cut into more than half of platform workers’ earnings. This study surveyed 407 drivers and 236 food and grocery delivery workers in the San Francisco Bay Area. Drivers typically earned $360 per week after expenses ($900 before), and the median earnings of delivery workers were $224 per week after expenses ($500 before).[265] Similarly, New York City’s Taxi and Limousine Commission, which regulates rideshare services in the city, estimates that drivers spend approximately 46 percent of their total compensation from apps on work-related expenses. To earn the city’s $15 hourly minimum wage, the Commission found that gross pay must be a minimum $27.86 per hour.[266]

In Texas, most workers Human Rights Watch surveyed indicated relying on cars for platform work; only two workers reported using lower-cost options such as their bicycle or walking. To calculate hourly wages of platform workers, Human Rights Watch accounted for all their work-related expenses, deducting costs that an employer would normally cover, to make it comparable to the situation of an employee.

The workers who responded to Human Rights Watch’s survey reported spending nearly half of their earnings on such costs of production. The table on the following page provides a breakdown of the expenses workers reported, including car insurance, car payments, gas, tolls, parking fees, car maintenance, and a phone plan. This is a conservative estimate that omits some costs, such as repairs, vehicle storage costs, and vehicle price depreciation because of platform work.

We calculated reported expenses on a per hour worked basis. Some workers purchased or leased their car for platform work, while others used cars they already owned. Our estimates reflect the cost of operating the vehicles, which applies to part-time as well as full-time workers.

The largest expense for platform workers in Texas was gas. On average, the workers surveyed spent $100 on weekly fuel expenses, or $2.76 for each hour worked. At the time of the survey, the price of gas in Texas was about $3 per gallon.[267]

Click to expand Image

The expense data were reported directly by workers in the survey. To validate our findings, we compared the expenses reported by workers against the rate set by the Internal Revenue Service (IRS), the federal agency responsible for collecting federal taxes, administering the Internal Revenue Code, and enforcing tax laws. The IRS rate for estimated expenses based on miles driven is generally considered to be relatively low, based on an annual study of the fixed and variable costs of operating an automobile, including car lease payments, maintenance and repairs, gas, oil, insurance, and vehicle registration fees.[268] For the period from July 1, 2022, to December 31, 2022, when the survey was conducted, the IRS rate was 62.5 cents per mile.[269] Applying the IRS rate to the average miles traveled for platform work yielded an average hourly expense of $9.37, which is higher than the expenses reported by workers, likely because workers omitted to share some of the maintenance costs.

Subtracting the IRS rate expense estimate from workers’ hourly remuneration yields an average hourly wage of $7.53.

One in five survey respondents reported expenses higher than their remuneration in the past week worked, even though we did not even ask them for all of their expenses. These findings resonate with research conducted in other parts of the country. For example, Benner et al. found in 2020 that among San Francisco delivery drivers, up to 12 percent may be operating at a loss after accounting for expenses. Similarly, researchers at MIT found in 2018 that after expenses, between 4 and 8 percent of workers were losing money.[270]

Workers testified about the importance of maintenance costs, many of which are difficult to assess for the workers, given their uneven frequency and the multiple use of their instruments of work, their car, for personal and professional use. Lorena K. shared: “the wear and tear on my car just killed me … I made $500 to $700 with Uber a week [including tips], I know that seems like a lot, but if I was making that, that would mean a lot of hours, probably 70 to 80.”[271]

Some workers buy or rent cars specifically for platform work. Research by the University of California Los Angeles (UCLA) found that in Los Angeles, 36 percent of rideshare drivers either purchased or leased a car for their work.[272] In Seattle, a study by researchers from The New School found that 83 percent of full-time drivers purchased their car for rideshare services.[273] In New York, 80 percent of drivers acquired their vehicle to engage in platform work.[274]

In Texas, nearly 4 out of 10 survey respondents told Human Rights Watch that they specifically purchased a vehicle for platform work. Among these individuals, monthly car payments averaged $115.

Step 3 – Deducting Social Security and other Benefit Contributions

The third and final step in estimating hourly wages is to account for the social security and other benefit contributions that employers typically provide to their employees; benefits and rights that, if provided at all, are only partially guaranteed to platform workers in the US.

In the traditional employment model, employees typically receive a compensation package that includes both wage and nonwage benefits. In the US, these nonwage benefits include federally legally required social security contributions to:

• Social Security: this funds the old-age, survivors, and disability Insurance (OASDI) program, which provides benefits to retirees, people with disabilities, and survivors of deceased workers.

• Medicare: the federal health insurance program primarily for individuals aged 65 and older, as well as certain younger people with disabilities.

• Unemployment insurance: employers pay into the state and federal unemployment insurance programs.

Additionally, many employers provide contributions to social security programs that are either voluntary or legally mandated by state government. These benefits are part of the internationally recognized social protection floor.[275] The most common benefits identified by the US Bureau of Labor Statistics[276] that employers contribute to include:

• Workers’ compensation: this insurance provides financial and medical benefits to employees who are injured or become ill due to their jobs.

• Additional retirement plans: a 401(k) is a retirement savings plan by employers, allowing employees to save and invest a portion of their paycheck before taxes are taken out. Employers may choose to match a portion of employee contribution or provide direct contribution.

• Health insurance premiums: employers may offer group health insurance plans, which provide coverage for medical, dental, and vision care. While not federally mandated, many employers contribute a portion of the insurance premiums, reducing the cost for employees.

• Life and disability insurance: offering financial protection in the event of death or a disabling injury or illness. Employers often cover part or all of the premium costs for these plans.

• Paid family and medical leave: paid leave for employees to care for themselves or family members during illness or after childbirth or adoption.

• Supplemental pay: such as additional compensation in dangerous working conditions (hazard pay), for hours worked beyond the standard workweek (overtime pay).

In several states, some of these benefits have been established as entitlements and are therefore legal requirements for employers. For example, 13 states and the District of Columbia, but not Texas, have enacted mandatory paid family leave laws.[277] All US states except Texas require employers to carry workers’ compensation insurance. In Texas, coverage is optional for most businesses.[278]

Taking both legally required and voluntary benefit contributions into account is necessary to make meaningful wage comparisons between platform workers and employees. In addition, benefits such as workers’ compensation, health insurance, and paid leave are essential for upholding workers’ rights.

However, platform workers, who are considered their own employers, must cover these contributions from the remuneration they receive from the platform. This significantly reduces their effective gross hourly wages.

The Employer Cost of Employee Compensation series from the US Bureau of Labor Statistics (BLS) estimates the national average employer costs for benefits for private sector workers. In the first quarter of 2023, social security and benefit contributions accounted for approximately 29.5 percent of total remuneration for private sector workers, with the remaining 70.5 percent allocated to wages.[279] On average, private sector companies across the country allocated $40.79 per hour to employee remuneration. Of that amount, $28.76 (70.5 percent) went to wages, while $12.02 (29.5 percent) was allocated to benefits.

Click to expand Image

In Step 3, we deduct the average employer costs for benefits from the remuneration after costs of production. For platform workers Human Rights Watch surveyed in Texas—who cover both mandatory and some voluntary benefit contributions from their hourly remuneration—this results in an average hourly wage of $5.12. This figure is derived by subtracting the total benefits cost of $2.41 from the hourly post-production cost remuneration of $7.53.[280]

Click to expand Image

This analysis reveals a stark disparity between platform workers’ wages and both minimum and living wage standards. With an average hourly wage of $5.12, platform workers surveyed in Texas earned about 29 percent less than the minimum wage of $7.25. As noted earlier, the minimum wage does not apply to platform workers because they are currently considered independent contractors rather than employees.

The discrepancy is even more pronounced when compared to the living wage threshold of $16.75 for a single adult, as platform workers earn at least 70 percent below this standard. Since the living wage is calculated after income taxes, and the provided estimate for platform workers is before tax deductions, the actual gap is likely even wider.

These figures illustrate the financial hardships experienced by many platform workers, as detailed in interviews with Human Rights Watch. They also underscore the urgent need for regulatory reform and legal protections in line with internationally agreed human rights standards.

Denied Social Protection

Strong social protection policies are powerful tools to reduce poverty and inequality and realize the internationally recognized right to social security and other human rights.

The human right to social security can be ensured through a range of programs, whether funded by direct contributions from workers and employers or through general taxation, that shield individuals from income insecurity, premised on the principle that everyone should enjoy their economic, social and cultural rights at every stage of their lives.

In the United States, the term “social security” has been largely conflated with a specific program that provides financial support to retired workers, people with disabilities, and survivors of deceased workers through monthly benefits. As used in international human rights instruments, social security encompasses nine areas of support: health care, sickness, old age, unemployment, employment injury, family and child support, maternity, disability, and survivors and orphans.[281] While the terms welfare and social safety net are frequently used in the US, there is no uniformly accepted definition for what they entail.[282]

The Universal Declaration of Human Rights affirms that everyone has the right to social security as a member of society.[283] Under the International Covenant on Economic, Social and Cultural Rights (ICESCR), ratified by 171 countries but not the US, governments undertake to respect, protect and fulfill the right to social security and are accountable for its realization.[284]

The UN Committee on Economic, Social and Cultural Rights, the treaty body that monitors implementation of the ICESCR, has specified that “steps must be taken by States parties to the maximum of their available resources to ensure that the social security systems cover workers inadequately protected by social security, including part-time workers, casual workers, the self-employed and homeworkers.”[285]

The United States has signed, but not ratified, the ICESCR. As a signatory, the US is obligated to refrain from acts that would defeat the treaty’s object and purpose.[286] As a state party to the International Convention on the Elimination of All Forms of Racial Discrimination (ICERD), the United States has also committed to “undertake to prohibit and to eliminate racial discrimination in all its forms and to guarantee the right of everyone, without distinction as to race, colour, or national or ethnic origin, to equality before the law, notably in the enjoyment of […] the right to public health, medical care, social security and social services.”[287]

A 2021 survey by the Pew Research Center found that 46 percent of platform workers in the US consider the level of benefits provided by the companies they work for as “unfair”.[288]

Social security systems typically comprise programs financed through general taxation, and programs financed primarily via direct contributions made by employers and workers.

Most informal workers who do not qualify for direct contributory programs through their employers may opt for private insurance mechanisms, which are entirely financed by the worker and generally result in higher premium contributions. In the US, the administration and regulation of these programs can vary, with some being coordinated and managed at the federal level, while others are governed by state-specific policies.

Exclusionary policies, administrative hurdles, and high costs for private alternatives are among the factors that prevent platform workers from accessing social security, as described in the following sections.

Exclusionary Policies

In Texas, platform workers are not eligible to contribute to or receive benefits from several social security programs available to employees, including unemployment and employment injury insurance; two programs that are mandatory for employees. This exclusion denies them essential protections that international standards call for, including support during periods of unemployment, compensation for work-related injuries, and access to family or medical leave during illness or while caring for a loved one.[289] This situation increases the risk of financial precarity and instability, posing a significant challenge to their overall well-being.

While platform workers in Texas are eligible for the two federally mandated contributory programs, Social Security and Medicare, they are required to cover both the employer and the employee shares of the contribution. Social Security and Medicare primarily provide benefits to individuals 65 and older, though some coverage is available for individuals with certain disabilities. Immediate healthcare and other benefits are generally not provided under these programs.
 

Disparities in coverage between traditional employees and platform workers across major social security programs in Texas

The International Labour Organization (ILO) has called on governments to ensure adequate social security protections “for all workers,” including those who are self-employed or incorrectly classified as self-employed.[298]

While some states have taken steps to expand contributory programs to platform workers,[299] Texas actively prevents platform workers from accessing certain protections.[300] The Texas Workforce Commission (TWC), the agency responsible for overseeing state workplace regulations, adopted a rule in 2019 amending the Texas Administrative Code (Section 815.134, Chapter 40), designating platform workers as “marketplace contractors” and ineligible for unemployment insurance.[301]

The US is the only high-income country among the Organization for Economic Co-operation and Development (OECD) that lacks federally mandated paid sick time or paid family leave.[302] Instead, the amount of paid sick and family leave is left to the discretion of employers. Paid sick leave allows workers to take time off without loss of pay to address their health needs. Workers without paid leave are more likely to continue working when ill and forgo medical care.[303] Paid family leave, meanwhile, allows workers to take time to support family caregiving needs, such as after the birth or adoption of a child or when navigating a loved one’s serious illness. Lack of paid family leave has significant consequences for maternal and child health, as well as families’ economic security.[304]

In the absence of a federal paid leave policy, some states have extended the option for independent contractors to opt into paid family and medical leave programs.[305] In New York state, a 2023 settlement mandated that even though platform workers are not classified as employees, Uber and Lyft drivers must provide their drivers with guaranteed paid sick leave. Drivers will earn one hour’s sick leave for every 30 hours worked, accruing up to a maximum of 56 hours’ sick leave per year.[306] Texas does not have paid leave provisions for platform workers.

In 2019, New York state mandated that all drivers for rideshare companies be covered by workers’ compensation insurance. This means that a rideshare driver who is injured in the course of their work is eligible for workers’ compensation benefits in New York.[307] In Texas, platform workers are not eligible for, and the platform companies do not contribute to, the state’s worker compensation fund.[308]

Administrative Hurdles

A second barrier preventing platform workers from accessing social security is the presence of administrative hurdles in programs for which they are eligible. The Social Security program for older age covers both employees and independent contractors, provided they make the necessary contributions, known as payroll taxes, in the US. Traditional employees make their share of contributions through automatic payroll deductions, with employers covering half the contributions. In contrast, platform workers face a more complex process. They are responsible for both the employee and employer shares of the Social Security contributions.

Platform workers may also qualify for most federal social assistance programs, such as the food assistance Supplemental Nutrition Assistance Program (SNAP), if their income falls below a specified threshold. These programs are typically financed through general taxation and provided independent of employment status.

However, even though platform workers can access programs like SNAP, proving eligibility can be difficult with fluctuating incomes. For SNAP eligibility, a household’s income should be at or below 130 percent of the federal poverty line, which was $36,075 for a family of four in 2022.[309] Applicants are required to submit documentation of household earnings, usually covering four weeks of pay stubs. While this verification process may pose minimal challenges for individuals with regular working hours and a consistent pay schedule, platform workers who experience irregular working hours, sporadic paychecks, and limited documentation of their earnings can face significant hurdles in the process.

Most platform workers do not receive pay stubs, and their hours per week may not accurately capture the nature of their employment. While platforms may offer digital records of payments, the only formal documentation of income is the Form 1099 sent in January of a given year. Additionally, providing evidence of actual wages necessitates proof of work-related expenses, adding an extra administrative burden.[310]

In recent years, several countries, including Indonesia, Uruguay, France, and Spain, have introduced measures to expand social security to platform workers. In Indonesia, the government facilitated the registration and payment of social security contributions to expand coverage of work injury and death benefits to drivers for Gojek , the country’s largest ride-hailing platform. This initiative encouraged Gojek drivers to formalize their work by registering with the government agency, with social security contributions deducted directly from driver accounts by the platform.[311] In Brazil and Uruguay, the governments extended their monotax mechanisms to cover drivers working for apps, granting them access to sickness, maternity and disability benefits as well as old-age pensions. Platforms are required to include social security contributions in the service cost and transfer them to the social security agency.[312] In Spain, legislation now mandates employment injury insurance for workers in dependent self-employment situations.[313]

High Cost of Private Health Care Alternatives

Unlike most high-income countries, the US has neither a universally accessible public healthcare system nor universal health insurance coverage.[316] Rather, the majority of the country’s population[317] relies on health insurance provided through employers to mitigate the potentially catastrophic costs associated with accessing health care in the country.[318]

To foster this large and highly fragmented system of so-called employment-based health insurance, the US both mandates that certain employers provide health insurance to their full-time employees[319] and further provides these employers with subsidies in the form of tax exemptions, which are collectively worth hundreds of billions of dollars each year.[320] Independent contractors, however, are excluded from this employment-based health insurance system.[321]

Since platform workers are not eligible for employment-based health insurance, their only means of health care coverage, if they do not have insurance through another job or a family member, is public health insurance programs, provided their income falls below poverty thresholds. For those earning above the poverty threshold, their only option for securing health coverage is through private insurance. A 2021 Rideshare Drivers United survey of more than 500 drivers in California, 65 percent of whom identified as people of color, found that 29 percent of respondents utilized California’s state-administered version of Medicaid, the public health insurance program for low-income adults.[322]

To qualify for Medicaid in 2024, a single person without dependents would need to earn below $20,782 in most states.[323] Except for Wisconsin, in all of the 10 states that have not yet adopted the 2010 Patient Protection and Affordable Care Act’s “Medicaid expansion,” including Texas, no single-person households without dependents are eligible for this public health insurance program. More than 1.5 million uninsured adults fall into this coverage gap nationwide, but about 40 percent of them live in Texas.[324]

The 2021 survey by Rideshare Drivers United of California-based drivers found that 16 percent of those surveyed were not covered by any health insurance, double the rate of national uninsurance among the general population. Another 29 percent of those surveyed were eligible for Medicaid.[325] Surveys conducted with platform workers in New York City, Seattle, and San Francisco found similarly high rates of Medicaid eligibility among platform workers, with 30 to 40 percent of workers surveyed reporting earnings below the eligibility threshold. In New York, 16 percent lacked insurance, compared to 27 percent in Seattle and 21 percent in San Francisco.[326]

Among the workers Human Rights Watch surveyed in Texas, more than one-third reported lacking any form of health insurance. Smaller numbers reported eligibility for Medicaid or using a family member’s health insurance.[327]

Impacts and Solutions to the Lack of Social Security for Platform Workers in Texas

The gaps in social security coverage are doubly challenging for platform workers. On the one hand, the lack of protection leaves workers exposed to the uncertainties of the labor market, undermines their rights, exacerbates poverty and inequality, and jeopardizes the implicit social contract. On the other hand, gaps in social security coverage for workers in certain contractual arrangements can distort hiring decisions, weaken investments in education and skills, and undermine competition among companies.[328]

Disputes regarding the classification of workers and its impact on benefits have emerged in various states. In some instances, agreements have been successfully brokered to extend benefits currently excluded for platform workers. A notable example is in New York, where Uber and Lyft reached a settlement with the State Department of Labor in November 2023 to facilitate eligible platform workers in accessing unemployment benefits. Uber also committed to making quarterly payments to the State Unemployment Insurance Trust Fund, contributing to the funding of benefits for unemployed workers, along with a retroactive payment covering amounts owed since 2013.[329]

By contrast, Texas has not yet addressed the social security gap, as benefits remain tethered to traditional employment structures, thereby effectively denying platform workers their rightful entitlement to social security.

Challenges to Organizing

Another set of rights denied to platform workers because of their employment status is the right to organize and collectively bargain over wages, benefits, and working conditions. All states, including the United States, have a duty to ensure the right of each worker to form and join trade unions under international labor rights law and other international human rights laws.[330]

Article 22 of the International Covenant on Civil and Political Rights (ICCPR), which the United States ratified in 1992, recognizes the “right to freedom of association with others, including the right to form and join trade unions for the protection of his interests.”[331] Article 8 of the International Covenant on Economic, Social and Cultural Rights (ICESCR) also recognizes and elaborates on these rights.[332]

The right to unionize is also protected by the fundamental conventions of the International Labour Organization (ILO), in particular Convention No. 87 on Freedom of Association[333] and the Protection of the Right to Organize and Convention No. 98 on the Right to Organize and Collective Bargaining.[334] The United States is an ILO member and has a duty to abide by these conventions’ terms, even though it has not ratified them.[335] The ILO also considers the right to strike to be an “intrinsic corollary” of the right to freedom of association if it is utilized as a means of defending workers’ economic and social interests. It also has said that the right to strike is one of the “essential means” for workers to improve their working conditions.[336]

The US has not established laws and policies that enable platform workers to exercise these rights. The National Labor Relations Act (NLRA) protects workers who advocate for improvements on the job or seek to organize a union, but it only extends those protections to workers classified as employees. Platform workers classified as independent contractors are not protected by the NLRA, depriving them of collective bargaining protections and union recognition, which, in addition to being a right, have been proven to enhance job quality and wages, and reduce income and wealth disparities. These protection gaps also make it difficult for established unions to defend platform workers’ rights and strengthen their bargaining position in wage-setting.[337]

In 2021, the US House of Representatives passed the Protecting the Right to Organize Act (“the PRO Act”), which aims to address some of these inequities by recognizing platform workers as employees for the purposes of organizing and unionizing, thereby providing them with a pathway to form a legally recognized union.[338] The bill did not advance in the US Senate.[339] In 2023, the PRO Act was reintroduced as the Richard L. Trumka Protecting the Right to Organize Act, but it was not brought to a vote in either the House or the Senate during that congressional session.[340]

In the meantime, digital labor platforms have lobbied for regulation that weakens the rights of platform workers to bargain collectively for better working conditions.[341] In California, the industry-backed Prop 22 law, passed by public ballot in 2020, denies platform workers the ability to bargain collectively or become members of unions to represent their interests, unless these arrangements are approved by a seven-eighths majority vote in the state legislature or through another ballot initiative, an extremely difficult bar to clear, and one that a California appeals court later found to be unconstitutional.[342]

Despite existing legal hurdles, some US platform workers have tried to organize to collectively bargain for better wages, benefits, and working conditions. However, a key challenge to organizing lies in the very nature of the work itself. As workers are assigned and paid by an app to compete for highly individualized routes and tasks, they lack a natural gathering point, such as a break room or an office, to forge relationships of solidarity and trust.[343]

Some groups of platform workers have repurposed makeshift meeting points into sites of collective action. At the Washington, DC, National Airport, Uber drivers who frequent the parking lot designated for rideshare pickups have coordinated efforts to improve working conditions.[344] In New York City, food delivery workers pool their earnings to rent spaces in parking garages where they can rest, share meals, and recharge their electric bikes in between shifts.[345]

But such meeting points are not available to all platform workers, such as grocery delivery workers who spend most of their work time in grocery stores or on delivery routes. “Supermarkets are not good places to integrate,” said Willy Solis, who was a lead organizer with Gig Workers Collective at the time.[346] “You don’t really know whether somebody is a shopper or not. You have some groups that text message each other but outside of that everyone is pretty much on their own.”[347]

High turnover among platform workers puts further strain on organizing efforts. Some estimates indicate that the turnover rate can be as high as 500 percent per year.[348] Solis said that many workers who began organizing with the Collective in April 2020 have since left Shipt and Instacart. As a result, they are no longer as involved in the Collective’s activities, or have stopped participating entirely. “They become disillusioned with the whole idea that they are running their own business, and they can make a good living,” said Solis.[349] As of the time of publication, the Collective is no longer active and Solis himself has cut back significantly on platform work.[350] In 2023, Solis stepped away from platform work altogether, citing the financial instability it created.

Tyler A., a platform worker in Ohio, made a similar observation about rideshare drivers. “Turnover is a challenge in every labor effort, but uniquely intense for us. I’ve had meetings with promising people [for organizing] and people would get invested and then they would get another job because they felt that driving was too risky [during the pandemic] and they were on unemployment anyway,” Tyler said.[351]

These challenges have made social media critical to organizing efforts. In October 2021, the Collective launched the #DeleteInstacart campaign, urging customers to delete the app until the company took concrete steps to change “the genuinely inequitable manner in which it treats its shoppers,” such as adopting fairer pay and ratings standards.[352] Earlier that year, leaders of the #DeclineNow campaign against DoorDash mobilized 26,000 workers on Facebook to decline low-paying orders on the app and force its algorithm to pay them a decent wage.[353]

The Shipt List, a collective of Shipt shoppers, has also staged digital walkouts on at least two occasions: in April 2020, to claim their rights during the pandemic, such as a more consistent supply of personal protective equipment, hazard pay and paid leave; and in July and October 2020, to protest the rollout of opaque pay algorithms.[354]

Despite these milestones, labor activists told Human Rights Watch that they are facing an uphill battle securing the resources they need to perform the labor-intensive work of grassroots organizing. Nicole Moore, the President of Rideshare Drivers United, described the struggle:

We have so many great ideas but not a lot of resources to do it. It all comes down to money [for] hiring staff, hiring communications and legal specialists. Even the law work we do is volunteer. We have filed several amicus briefs. Volunteer lawyers who helped us file those. It’s great, but it’s twice as much work [to organize them] and you don’t always have resources to make that happen.[355]

Organizers trace many of these resource constraints to the misclassification of app-based workers as independent contractors, which means that they do not have a legally protected right to organize and form a union. “There is no doubt there would be more resources if there’s a clear path to unionization,” Moore said. “Without a clear pathway legally, there are few [established] unions investing in building real unions in this sector.”[356]

The lack of legal protections can also lead to workers refraining from engaging in organizing activities due to fear of retaliation. The way that platforms treat some of their employees may be contributing to these fears. In 2021, Instacart laid off every one of the ten in-store employee shoppers who had formed the company’s first and only union, as part of a broader layoff of 2,000 in-store shoppers.[357] The shoppers had formed the union only a year before.[358] The company characterized these layoffs as part of a shift in its business strategy.[359] In July 2023, the National Labor Relations Board filed a complaint against DoorDash alleging that it had illegally prohibited its office employees in Arizona from discussing working conditions during their days off and fired one for attempting to organize.[360] DoorDash said that these allegations “are part of a personally motivated attack and entirely without merit.”[361]

Some platforms have tried to form alternatives to unions. In 2016, Uber entered into a five-year agreement with the International Association of Machinists and Aerospace Workers to fund and create an “Independent Drivers Guild”, on the proviso that it would create a space for workers, while completely lacking in necessary independence from Uber management.[362] Although the agreement established monthly meetings with Uber management and a process for drivers to appeal deactivations, it did not permit the Guild to bargain collectively on behalf of drivers. The Guild would also refrain from encouraging drivers to strike or challenging Uber’s classification of drivers as independent contractors, even though individual drivers were free to challenge it on their own.[363] The Guild, which was first established in New York, now has chapters in New Jersey, Massachusetts, Illinois and Connecticut.[364] Uber and Lyft proposed a similar arrangement in California before the passage of AB5.[365]

Both international and US labor standards recognize that genuine worker representation cannot exist when unions are subject to employer control. Article 2 of ILO Convention 98 on the Right to Organise and Collective Bargaining Convention prohibits “acts which are designed to promote the establishment of workers' organisations under the domination of employers or employers' organisations, or to support workers' organisations by financial or other means, with the object of placing such organisations under the control of employers or employers' organisations.”[366]

Similarly, under US labor law, section 8(a)(2) of the NLRA explicitly prohibits employers from dominating or interfering with the formation or administration of any labor organization. This means that employer-created or influenced unions, referred to as "company unions," are considered an unfair labor practice. However, the application of this protection hinges on the classification of workers as employees under the NLRA, meaning that independent contractors and others not recognized as employees may not be covered by these provisions.[367]

A Labor Model That Traps People in Poverty

The combination of black box algorithms surveilling, controlling, and deactivating workers, along with low and unpredictable pay, lack of social security, and obstacles to unionizing translate into a system that lends itself to exploitation. These conditions can trap workers in a constant state of economic insecurity.[368]

Academics, labor economists, and practitioners have highlighted the broader implications of this trend on economic inequality in the US.[369] In a 2023 report titled “The working poor,” the United Nation’s Special Rapporteur on extreme poverty and human rights, Olivier De Schutter, alleged major US-based platform companies systemically underpaid their platform workers.[370] He stated:

I am extremely disturbed that workers in some of the world's most profitable companies – in one of the richest countries on earth – are struggling to afford to eat or pay their rent. Multi-billion-dollar companies should be setting the standard for working conditions and wages, not violating the human rights of their workers by failing to pay them a decent wage.[371]

The existing US regulatory framework, which allows the misclassification of workers as independent contractors rather than employees, is at the heart of the problem. It allows companies to bypass requisite taxes and avoid paying for the employment benefits and protections that formal employment would require. In effect, the bulk of employment costs, including programs to ensure workers’ rights like social security, are shifted onto the workers themselves and society at large. This shift enables companies to increase their profit margins to the detriment of workers’ rights and the public interest.

Technology as a Veneer for Worker Exploitation

Platform companies often justify this misclassification by insisting they are technology companies, not transportation or delivery companies, and therefore not subject to the same legal obligations of employers and standard companies.[372] This “tech” framing creates a veneer of legitimacy and shields them from government scrutiny, while portraying their business model as “innovative” rather than exploitative.

Technology is intricately woven into the fabric of these companies’ operations, serving multiple functions: it facilitates the recruitment and maintenance of available drivers and delivery workers, enables customers to order services or products on demand, connects workers with orders, and dispatches them to fulfill jobs. The review systems perpetuate behavioral norms, and various incentives, as discussed earlier, further entrench control over workers. Furthermore, the design of these systems prevents operations from being scrutinized, achieved through algorithmic processing and via opaque black box decision-making. The use of technology also creates a barrier between platform workers and employers, inhibiting workers’ ability to have a say in how they are paid or to pursue career development within the company.

The increasing use of technology in labor markets would not be as concerning if earnings were more fairly distributed and workers were provided with a living wage that allowed them to fulfill their rights. But the current model risks entrenching existing economic inequality, by trapping workers in a system that marginalizes them. This labor model denies platform workers internationally recognized rights, including the rights to a minimum wage and to social security. Coupled with unpredictable pay and schedules, platform workers face daily struggles to maintain an adequate standard of living and enjoy their rights to health, education, rest, and cultural life.[373]

As unanimously recognized in the ILO 1944 Declaration of Philadelphia, reducing inequalities requires adequate protection of all workers and ensuring that they benefit from a “fair share of the fruits of progress.”[374] Fundamental principles and rights at work should be guaranteed to all workers, regardless of their employment relationship. Collective bargaining and measures for equal pay for work of equal value and the enforcement of pay transparency are also key for achieving fair distribution of work. Adequate minimum wages, statutory or negotiated, are also a powerful tool to reduce inequalities and should be set by taking into account the rights of workers and their families, as well as economic factors, as laid out in ILO Convention No. 131.[375]

Interviews conducted by Human Rights Watch with platform workers across the US shed light on the financial insecurity many endure, making it challenging to pay for items that are rights such as adequate food or housing. Some workers accumulate debt or are forced to pawn personal items to make ends meet, while others reported feeling trapped in this line of work. John R., who primarily works for DoorDash in Dallas, said: “I just got a new car last week, and I already put 3,000 miles on it. […] Once you take on more financial responsibilities, you'll have to be more open to the app work […] [I have] higher insurance since it's a newer vehicle, and I'm paying more for my car note.”[376] Despite earning roughly $500 per week through app-based work, he allocated over 80 percent to cover work-related expenses.[377]

This reality underscores the urgent need for appropriate regulations across the platform work industry in the US. Without robust safeguards and state oversight, technology, rather than improving working conditions, can become a tool to exploit workers, perpetuating a cycle of financial instability and inequality.

The growth in revenue and market share that several platforms have recorded in recent years has not translated into living wages for platform workers, leaving many disillusioned. Mae D., a Dasher in Portland, said platform work reaps “an enormous profit for DoorDash and an enormous transfer of wealth out of the pockets of low-income Americans who are really just trying to find a way to continue to make ends meet."[378] Frankie C., a Shipt shopper in Dallas, echoed the sentiment, expressing demoralization as corporations celebrate financial successes while workers struggle. “They are releasing numbers: ‘Oh, we’ve made so much percent more than we’ve ever made forever.’ And it’s like, really, are you sure? Because we are not. It’s not reflecting on any of us.”[379] In 2021, the labor groups Gig Workers Rising and We Drive Progress organized a petition and protested in front of the house of the DoorDash CEO, demanding “share the wealth.”[380]

Research by the ILO and the UK’s Economic and Social Research Council found that inequality is at the core of platform work, arising from various asymmetries, including of financial benefits, risk, information, and power.[381] These studies suggest that financial gains primarily accrue to platform companies and customers, leaving workers with substantial risks due to the absence of labor standards and social security.[382] Marshall Steinbaum, Assistant Professor of Economics at the University of Utah, argues that barriers preventing platform workers from collective bargaining further exacerbate this inequality.[383]

Since the 1930s, the Securities and Exchange Commission (SEC), an independent federal agency, has used its regulatory power under the securities laws to compel the disclosure of certain pay information by public companies.[384] In 2015, the SEC adopted a new pay ratio rule, a provision in the Dodd-Frank Act, a financial regulation law enacted in 2010.[385] This rule mandates that public companies disclose the median earnings of all employees, including full-time, part-time, temporary, and seasonal workers. Additionally, companies are required to publish a ratio comparing the compensation of their highest earner to that of their median employee in annual proxy statements. However, the rule excludes individuals employed by unaffiliated third-parties or independent contractors. If platform workers were correctly classified as employees, public companies would be obligated to report their median earnings, ensuring more transparency and accountability in platform work.

The Public Cost of Unmet Social Security Contributions

In the US, the 1935 Social Security Act and the Federal Insurance Contributions Act (FICA) mandate that employers contribute to the federally mandated social security programs—Social Security, Medicare, and unemployment—via so-called payroll taxes. By classifying platform workers as independent contractors, companies avoid these contributions that are required of other employers. Instead, workers must cover both the employee and employer shares of Social Security and Medicare contributions. They are also ineligible for unemployment benefits and cannot contribute to state unemployment funds, resulting in forgone tax revenues that states could otherwise collect if workers were properly classified.

A social security system that covers only specific categories of workers and exempts some corporations from contributing undermines risk pooling. Risk pooling in social security involves spreading financial risks among a large number of participants, ensuring that the cost of benefits is shared by everyone, thereby reducing the burden on any single individual. When coverage is limited to specific worker categories and not all corporations contribute, the pool of contributors shrinks. This reduction can lead to higher costs for those covered and can compromise the system’s stability and sustainability.

While platform firms gain a dual cost advantage by avoiding social security contributions and not having to pay employees during down-times, workers face a double risk by missing out on benefits such as unemployment benefits and paid sick leave. The cost is then de facto transferred to the public and other taxpayers, workers end up relying on non-contributory programs such as food stamps or Medicaid instead of insurance partly paid by the employer. As a result, taxpayers and state budgets are essentially subsidizing platform companies.

Several states have been challenging these practices due to the financial impact on their public funds. For example, a 2019 audit by New Jersey’s Department of Labor and Workforce Development found that due to the misclassification of drivers as contractors rather than employees, Uber owed $523 million in taxes for unemployment and disability insurance from 2014 to 2018. The state also demanded $119 million in interest.[386] In 2022, Uber settled by agreeing to pay New Jersey $100 million in back taxes; less than a fifth of what the audit had assessed as the company’s tax liability. [387]

Similarly, the New York State Department of Labor (NYSDOL) maintains that rideshare drivers and couriers are employees for purposes of unemployment insurance. In a 2023 settlement between the NYSDOL and Uber, the parties agreed that drivers and delivery workers eligible for unemployment benefits should receive them. It mandated that Uber start making regular payments to the state’s unemployment insurance program as well as retroactive payments dating back through 2013.[388] The governor’s office did not disclose the total amount of the settlement, citing confidentiality laws on unemployment insurance data.[389]

The University of California, Berkeley Labor Center estimates that if Uber and Lyft had classified workers as employees, the companies would owe $413 million to California's Unemployment Insurance (UI) fund for 2014 to 2019.[390]

To assess the magnitude of this issue in Texas, Human Rights Watch conducted an analysis using both microdata from the American Community Survey (ACS) and tax-filing-based data from the Census Bureau’s Nonemployer Statistics (NES), following methodologies developed by the UC Berkeley Labor Center and the Center for New York City Affairs.

Based on NES data, which captures self-employment income from both primary and supplemental work, Human Rights Watch estimates that Texas lost more than $111 million in forgone Unemployment Insurance (UI) contributions between 2020 and 2022 due to the misclassification of platform workers. This estimate includes individuals performing rideshare, delivery, and in-home labor services such as cleaning and repairs. It is derived from the number of workers (self-proprietors) reporting income in relevant sectors, their average annual receipts, and a modeled income distribution relative to the state’s UI taxable wage base of $9,000. In 2022 alone, more than 429,000 individuals in Texas filed as self-employed in platform related industries (see Methodological Note Box below).[391]

Click to expand Image

Forgone contributions risk the further weakening of an already fragmented social protection system. They can deplete state insurance funds, leading to increased reliance on federal loans to cover benefits. This situation contributes to federal debt in unemployment insurance funds and can undermine the stability and adequacy of the social protection system.[392]
 

Because platform companies are not currently required to report how many workers they engage, how many hours they work, or how much they earn, available data is limited. The Nonemployer Statistics (NES) dataset—while valuable—misses workers who do not file tax returns, including those with very low earnings, informal work arrangements, or unreported income. As a result, the true scale of platform work and associated tax losses are almost certainly understated.
 

The Way Forward

The coming years will be pivotal in defining decent work within the platform economy. The International Labour Organization (ILO) Governing Body has announced that its 2025 Conference will include an agenda item dedicated to discussing standards related to decent work in the platform economy.[395] This development opens the door for the ILO to advance discussions and, ideally, approve an international convention that establishes guidelines ensuring fair work standards for the millions of workers providing services via digital labor platforms globally.

In addition to the government officials and unions advocating for this agenda within the ILO, over 300 professors, lecturers, and researchers launched a Global Manifesto in 2023 for a Fairer Platform Economy.[396] The manifesto calls for the approval of an international convention to secure decent work standards for platform workers.[397]

This report contributes to these discussions, documenting the ways in which platform companies have exploited the margins of the law to create a new labor model. This new model erodes the rights of US workers, contributes to their exploitation, and perpetuates patterns of inequality. Digital labor platforms can and should take immediate steps to uphold their responsibility to respect workers’ rights, which are detailed below. But fully addressing these violations requires a broader reimagination of platform work, beginning with regulation that rebalances the power between platforms and workers. This involves a multi-pronged, whole-of-government effort to shore up labor and privacy protections, impose comprehensive guardrails on AI in the workplace, and establish a universal and adequate social security system for all workers. These reforms should also be tailored to address historical racial and gender-based economic disparities that are endemic in the US labor market, and deepened in platform work.[398] In particular, policymakers and regulators should prioritize the following reforms:

1. Proper employment classification

2. Protection from algorithmic exploitation

3. Guarantee of fair and living wages

4. Provision of universal and adequate social security

5. Protection of worker agency and bargaining power

Proper Employment Classification

The US should urgently develop standards and policies that ensure platform workers' proper classification as employees. These should heed the ILO’s recommendation against “disguised employment relationships”:[408]

National policy should at least include measures to … combat disguised employment relationships in the context of, for example, other relationships that may include the use of other forms of contractual arrangements that hide the true legal status, noting that a disguised employment relationship occurs when the employer treats an individual as other than an employee in a manner that hides his or her true legal status as an employee, and that situations can arise where contractual arrangements have the effect of depriving workers of the protection they are due.

The “ABC” test is an employment classification standard that would be consistent with this recommendation.[409] This test shifts the burden to employers to demonstrate that they have properly classified their workers as independent contractors. The worker is presumed to be an employee unless their employment meets the following criteria:

• (a) Control: the worker operates independently, free from employer control,
• (b) Scope of work: the worker’s tasks fall outside the usual course of the employer’s business, and
• (c) Independent business: the worker is engaged in an independently established business.

Federal, state and local authorities that adopt the ABC test or a comparable classification standard would ensure that these workers are covered by applicable minimum wage and employment protections when working for platforms. Using such a standard would also ensure that platforms pay the required amount of social security and other benefit contributions to protect these rights for workers.

Platforms studied in this report—and any similar platform engaging workers in comparable ways—should voluntarily classify their platform workers as employees. They should also refrain from labor practices that obscure their true employment status, such as agreements that require workers to affirm that they are independent contractors as a condition of working for the platform.

The US Department of Labor could also clarify that platform workers should be classified as employees under its rules, which would make them eligible for protections under the Fair Labor Standards Act (FLSA). This would ensure that platform workers receive protections such as minimum wage, overtime pay, and other protections that are currently guaranteed to employees but often denied to independent contractors. Congress should also amend the definition of an employee under federal workplace discrimination laws, such as Title VII of the Civil Rights Act, to ensure that platform workers are covered. Without this amendment, platform workers do not have the same protections against discrimination that employees receive.[410]

Algorithmic Fairness

Rectifying the misuse of algorithms to depress wages, obscure true earnings, and penalize workers will require the US to pair strong privacy protections for workers with robust safeguards against algorithmic decision-making that negatively affects working conditions.

Protecting Workers’ Privacy

Protecting the privacy of platform workers begins with clear limits on collecting their personal data. The international privacy organization Privacy International has proposed restricting such collection to data that is “necessary for the evaluation or attribution of work,” and “only during periods agreed by workers”.[411] Under this standard, platforms should be prohibited from collecting workers’ geolocation and device usage data when they are not working or seeking work.

Abuses of workers’ personal data are frequently difficult to challenge because they are shrouded in secrecy. Workers should be granted the right to request information about the existence of automated decision-making that uses their personal data, how such decision-making works, and how it affects them.[412] This will also enable workers to contest automated decisions about them that are made without meaningful human intervention and other protective measures.[413]

The US should also establish a strong federal data protection law that empowers individuals to control how their personal data is used after they hand it over to an employer, company, or a government.[414] Many of the opaque and exploitative practices documented in this report, such as dynamic wage setting algorithms and targeted rewards schemes, are only possible because of access to workers’ data that platforms may collect for legitimate purposes. For example, the geolocation data used to dispatch work requests or track progress can also be used to target and personalize wage and bonus offers.

To prevent the abuse of workers’ personal data, federal and state authorities should use legitimately collected data from platform workers only for work allocation and performance management functions that do not pose an unacceptable risk to workers’ rights. The European Union’s General Data Protection Regulation (GDPR), for example, requires personal data to be “collected for specified, explicit and legitimate purposes and not further processed in a manner that is incompatible with those purposes.” The GDPR’s data minimization approach is in line with human rights principles of necessity and proportionality.[415] While privacy experts have indicated that the GDPR’s purpose limitation provisions allow for problematic exceptions, it provides a starting point for meaningful regulation.[416]

The European Union’s platform work directive (2024) places clear limitations on platforms’ ability to access, process, and store workers’ personal data. In addition to constraining platforms’ ability to process workers’ biometric data for identification purposes, the directive states that:

(40) Digital labour platforms should not, by means of automated monitoring systems and by means of any automated system used to support or take decisions affecting persons performing platform work, process any personal data on the emotional or psychological state of the person performing platform work, process any personal data in relation to their private conversations, collect any personal data while the person performing platform work is not offering or performing platform work, process any personal data to predict the exercise of fundamental rights, including the right of association, the right of collective bargaining and action or the right to information and consultation, as defined in the Charter, and they should not process personal data to infer racial or ethnic origin, migration status, political opinions, religious or philosophical beliefs, disability, state of health, including chronic disease or HIV status, the emotional or psychological state, trade union membership, a person's sex life or sexual orientation. [417]

The GDPR and the platform work directive both contain provisions for transparency for workers around automation. The GDPR allows data subjects to request relevant information around automation, and to contest where they are impacted by automated decision-making; the platform work directive states that platforms must communicate to workers in an accessible format regarding where and how automation is used to manage their work.[418]

AI-Related Safeguards

Rights-based regulation of algorithmic management practices should govern the entire life cycle of the algorithmic system, from data collection and use in the development of the system, to deployment and modification. Such regulation should, at a minimum, ban algorithmic management practices that pose an unacceptable risk to workers’ rights. For systems that create manageable risks, the regulation should impose mandatory impact assessments, audits, accountability for misuse of these systems, and remedies for workers harmed by them.  

Red Lines

Human Rights Watch’s research identifies three algorithmic management practices that should be banned altogether: 1) dynamic wage-setting algorithms; 2) ratings-based work allocation; and 3) fully automated deactivations.

Dynamic wage setting algorithms: Every example that Human Rights Watch came across of a dynamic pricing-style algorithm used by a platform to determine pricing had the effect of depressing wages for a sizable proportion of their workers. Such systems also generate arbitrary pay fluctuations and confusion about true earnings. It is also doubtful that existing privacy and AI-related safeguards, such as transparency requirements and limitations on data use, will effectively mitigate the effects of these algorithms. Professor Veena Dubal, a technology and labor expert, has argued that these inadequacies justify a “nonwaivable ban on algorithmic wage discrimination, including… a ban on compensation through digitalized piece pay.” She explains:

While knowing generally how the algorithm works might diminish the feeling of being manipulated, given the rapid rate at which machine learning systems change compared to the temporal tendencies of legal requests and subsequent adjudication, this knowledge will have little impact on drivers’ ability to exert control on the job or on wage standardization in a fair and predictable way.[419]

Ratings-based work allocation: Platforms have failed to sufficiently demonstrate that using customer ratings as a metric for assigning work leads to fair compensation for workers, based on an accurate and non-discriminatory assessment of their performance.[420] Human Rights Watch has also found that the remedial mechanisms that are supposed to enable workers to challenge unfair or fraudulent ratings often do not provide effective redress. Even when unfair or fraudulent ratings are reversed, platforms do not mitigate the resulting negative impact on workers’ incomes. Digital labor platforms should be required to decouple ratings systems from decisions about how to assign work.

Fully automated deactivations: Human Rights Watch has found that deactivating workers’ accounts temporarily or permanently has a significant negative impact on their ability to enjoy their human rights, in particular their rights to an adequate standard of living and a decent living.[421] Even if workers’ accounts are restored, the temporary loss of earnings has knock-on effects on their ability to cover housing costs, put food on the table, and provide for their children and families. Such consequential decisions should not be delegated entirely to automated systems. Automation-assisted deactivation systems should be subject to strict safeguards, which are detailed below.

Impact Assessments

Algorithmic impact assessments have become a cornerstone of human rights due diligence in the technology sector.[422] These assessments should be mandatory, transparent, and conducted in consultation with impacted workers. Platform companies should not be permitted to introduce or modify an algorithm that significantly affects working conditions, unless the company shows that it can effectively mitigate the risks these changes pose to workers’ rights. It is not clear whether any of the companies in this report have conducted them. Human Rights Watch asked each of the seven companies if they conducted algorithmic impact assessments or other relevant risk assessments to calculate the impact of algorithmic systems on working conditions, and to share relevant documentation, but did not receive an answer to this question from any of the companies.

Platform algorithms that significantly affect working conditions should also be subject to a regular program of audits throughout their lifecycle. The Ada Lovelace Institute, an artificial intelligence policy research center, sets out two broad categories of auditing practices: 1) technical audits, conducted by the developer (in this case, the platform company) or an independent third-party that investigates an algorithm for bias or error; and 2) regulatory audits conducted by the government that investigates an algorithm for compliance with applicable laws and policies, taking into account “the context [in which] it operates,” including “the organizational processes and human behavior around it.”[423]

Digital labor platforms should be required to conduct or commission regular technical audits of relevant algorithmic systems, such as their ratings and tipping systems, publish the audit’s findings, and take prompt and effective corrective action. These audits could have, for example, proactively identified and rectified glitches in Instacart and Shipt’s tipping systems that erroneously withheld tips from shoppers in 2018 and 2020 respectively.[424]

A handful of federal agencies appear well-positioned to conduct regulatory audits of platform algorithms. The Federal Trade Commission, which enforces antitrust and consumer protection laws, has signaled that it is scrutinizing claims of “potential earnings” and promises of a “steady second income” advertised by digital labor platforms, a possible basis for an investigation into how opaque wage-setting algorithms and the gamification of bonus schemes mislead workers about their true earnings and induce them to work long hours for low pay.[425] The Equal Employment Opportunity Commission, which enforces workplace discrimination laws, could invoke its authority to conduct technical studies under Title VII to investigate whether the algorithmic systems that digital labor platforms use to assign, pay, and rate work have a disparate impact on the earnings of women, people of color, workers with disabilities, and migrants.[426]

Regulating complex technology in a growing industry requires a concerted, multi-agency approach that brings technical, policy and labor expertise to bear on algorithmic systems that are changing the nature of work. Alex Engler, a Fellow in Governance Studies at the Brookings Institute, has urged the White House to ramp up hiring of data scientists and engineers to support the capacity of federal agencies to audit algorithmic systems.[427] Experts with inter-disciplinary experience, particularly in sociology, technology, law, and policy, would also be needed for such audits.

Effective remedies

Platform workers should have access to an effective remedy when they are unfairly fired or penalized by algorithmic systems. To protect workers from unfair deactivations, the federal and state governments should recognize the rights of workers to obtain a meaningful explanation of a decision to deactivate and to challenge such decisions, regardless of their employment status. These rights should apply to deactivations of any duration.

To meaningfully challenge unfair deactivations, workers must first understand them. Workers should be provided with a written statement clearly explaining the reasons for their deactivation, and the process and timelines for appealing the deactivation. They should also have the right to request the platform to conduct a human review of the deactivation, and a written response providing reasons for affirming or reversing the decision within a reasonable time.

To help ensure accountability and oversight, workers should be able to appeal deactivations that the platform has upheld after internal review to an independent third party. In 2023, for example, the state of Washington established a process for appealing decisions to deactivate rideshare drivers to an independent dispute resolution mechanism, and a Driver Resolution Center that provides drivers with representation in these proceedings.[428] But this process offers only limited protections: it exempts certain types of deactivations from third party review, such as those based on allegations of fraud, discrimination and harassment, and does not cover food and grocery delivery workers. A legislative proposal to create a rideshare driver bill of rights in Massachusetts, as well as the EU Platform Work Directive, would also establish mechanisms for challenging deactivations.[429]

When the platform or an independent appeals mechanism has found that a worker’s account has been deactivated unfairly or in error, the platform should reinstate the worker’s account immediately, and compensate workers for lost wages during the period of deactivation.  

Guarantee of Fair and Living Wages

The US Department of Labor should extend the Fair Labor Standards Act coverage to all platform workers by classifying them as employees to guarantee they earn at least a living wage. Previously, certain extensions of FLSA coverage have been granted to specific categories of workers, such as domestic service workers, under certain conditions.[430] In 2022, the Department of Labor proposed to modify Wage and Hour Division regulations to revise its analysis for determining employee or independent contractor classification under the Fair Labor Standards Act (FLSA) to be more consistent with judicial precedent and FLSA’s text and purpose. The DOL should adopt the proposed rule to reduce risk of misclassification as independent contractors.[431]

The federal minimum wage falls below a fair wage necessary for an adequate standard of living. The federal minimum wage has been US$7.25 per hour since 2009.[432] Numerous studies have indicated that this rate is insufficient to sustain an adequate standard of living, and if kept up with inflation should have been $23 per hour in 2021.[433] Congress should determine a living wage standard and adjust the federal minimum wage accordingly, ensuring it is automatically indexed to inflation moving forward.[434]

What is considered a living wage varies between countries and localities. According to MIT, it should consider the following costs, after their subsidies and direct provision by the state: food, childcare, healthcare, housing, civic engagement, broadband, transportation, income and payroll taxes, and other necessities.[435]

The wages provided should also be transparent. In addition to refraining from dynamic wage setting algorithms, platforms should ensure that the pay formulas they adopt enable workers to clearly understand how they are paid and how those wages are calculated. Workers’ rights advocates have, for example, called for grocery delivery platforms to return to a commission-based model of compensation, as this accurately reflects the time and effort invested in each order.[436]

Provision of Universal and Adequate Social Protection

The federal and state governments should guarantee that all platform workers are covered by adequate social security programs that ensure financial security during life course contingencies, including income security during injury, illness, job loss, older age or when caring for a loved one, as well as in the event of disability and survivorship, maternity, and for families with children. This requires the development of rights-aligned policy, legal, and administrative frameworks that adapt to the dynamic nature of platform work.

First and foremost, platform workers should be accurately classified as employees. Proper classification would not only grant immediate access to their social security and other labor rights, but it would also strengthen the overall social security system. By addressing worker misclassification, platform companies would be required to contribute their share to social security, including unemployment and insurance. This would bolster state social security resources, enhance protection for all workers by expanding the insurance pool, and reduce the financial risks across the workforce.

In the interim, as misclassification issues are being resolved, social security programs currently limited to employees should be expanded to include self-employed workers, including those with low incomes. Ensuring contributory programs are accessible to all workers, regardless of their employment status, would protect the rights of platform workers and benefit the broader population of informal workers.[437]

In terms of concrete legislation, the Department of Labor should expand paid sick and paid family leave programs to cover platform workers. In Texas, the proposed Family And Medical Insurance Leave (FAMILY) Act would provide all workers with up to 12 weeks of partial income when they take time for their own health or to care for a dependent.[438] The cost for the program would be spread across employers and workers.[439] The National Partnership for Women and Families estimates that the Act would reduce by 79 percent the number of working families in Texas facing significant economic insecurity when taking family and medical leave.[440]

Additionally, the Department of Labor (DOL) should provide regulatory guidance and policy recommendation to expand eligibility criteria to include platform workers. Congress should pass legislation that modifies existing law to extend coverage to these workers. States, guided by the principles of the ABC test, should ensure that all workers have unhindered access to unemployment benefits and that benefit levels adequately reflect the workers’ economic, social and cultural rights, including in particular the right to an adequate standard of living. The Texas Workforce Commission should retract the 2019 amendment to the Section 815.134 of the Administrative Code on Employment Status, which categorically excludes platform workers from unemployment insurance.[441]

The US should ensure that everyone has access to the highest attainable standard of health, regardless of their ability to pay.[442] To do so, US legislators should consider measures taken by many other high-income countries, such as creating a public healthcare system that is universally accessible for all, universal health insurance coverage, or some combination of these two. In recent years, states including Washington, California, New York, and New Jersey have experimented with policies that can guide national policy that ensures universal health care.[443]

While building toward a rights-aligned healthcare system, US authorities should require all applicable large employers to provide health insurance coverage for all workers that meet the principles of the ABC test. Additionally, states that have not implemented the so-called Medicaid expansion, which, among other things, expands the public health plan’s coverage to single persons without dependents, should also adopt legislation integrating these components of the program into state laws.

Another pathway to social security is facilitating benefits portability between jobs while ensuring continued protection. For example, in France, the unemployment insurance scheme was modified in 2014 to introduce “refillable rights.” Such schemes allow workers to accumulate benefits continuously, ensuring that transitions between jobs do not jeopardize their social protection.[444] In Washington state, proposed legislation (HB1601) would introduce a portable benefits fund financed by platform companies, enabling workers to access health insurance, retirement, and paid vacation benefits, irrespective of their employment classification.[445]

Protecting Workers’ Autonomy and Bargaining Power

The US should revise its existing legislation to grant platform workers the rights to freedom of association and collective bargaining, in accordance with international labor rights standards.

A pivotal step in this direction would be the passage of the federal Richard L. Trumka Protecting the Right to Organize Act (PRO Act), which would strengthen collective bargaining protections and broaden their coverage.[446] Specifically, it would extend to workers classified as employees under the ABC test. By passing the PRO Act, the US would get closer to meeting its obligations under international law by eliminating many of the barriers that prevent workers from exercising their rights to organize a union, extending bargaining protections to an estimated 1.6 million platform workers.[447]

In addition, alternative models of platform work can rebalance the power dynamics between workers and employers. Worker-owned platform cooperatives can give platform workers an ownership stake in companies and a share in profits. Rather than acting as a barrier between a company and workers, platform cooperatives serve as a common ground on which workers can offer services while connecting with one another. Owning the platform means owning the data and building wealth as the platform grows, reversing the pattern of exploitation apparent in the dominant models of today.[448]
 

International Human Rights Law

This report is about the human rights abuses of United States platform work. While the US has not yet ratified the International Covenant on Economic, Social, and Cultural Rights (ICESCR), as a signatory it is bound to refrain from actions conflicting with the treaty's object and purpose. The issues addressed in this report implicate a range of human rights protected under ICESCR and other international law standards, including the International Covenant on Civil and Political Rights (ICCPR), which the US has ratified.

These include the right to an adequate standard of living, the right to just and favorable conditions of work, the right to privacy, the right to safe and healthy working conditions, the right to social security, the right to health, and the right to unionize.

Right to an Adequate Standard of Living

The Universal Declaration of Human Rights (UDHR) states that everyone has a right to a “standard of living adequate for the health and well-being of himself and of his family, including food, clothing, housing and medical care and necessary social services.”[449] Article 11 of the ICESCR elaborates on this right.[450] The UN Committee on Economic, Social and Cultural Rights has interpreted the content and related obligations of the various components of the right as including the rights to housing, food, water, and sanitation.[451]

Right to Just and Favorable Conditions of Work

International human rights law protects a spectrum of workers' rights, and notably rights related to just and favorable conditions of work. The UDHR provides that everyone has the right to rest and leisure, including reasonable limitation of work hours and periodic holidays with pay.[452] The ICESCR reiterates these rights in recognizing the right of all persons to just and favorable conditions of work.[453]

Though a living wage is not nominally defined under international law, the UDHR states that “everyone who works has the right to just and favorable remuneration ensuring for himself and his family an existence worthy of human dignity, and supplemented, if necessary, by other means of social protection.”[454] The International Labour Organization sets forth this right in Conventions 95 and 131 and includes recommendations on enforcing minimum wages, including by protecting the right to freedom of association.

The ICESCR recognizes the right to a remuneration that provides for “a decent living for themselves and their families” in accordance with other basic rights protected by the Covenant and establishes “the right of everyone to the enjoyment of just and favorable conditions of work,” including “fair wages and equal remuneration for work of equal value,” and “a decent living for themselves and their families.”[455] 

According to the UN Committee on Economic, Social and Cultural Rights, which issues authoritative interpretations of the ICESCR, the right to just and favorable conditions of work requires a legally mandated minimum wage that is “non-discriminatory and non-derogable, fixed by taking into consideration relevant economic factors and indexed to the cost of living so as to ensure a decent living for workers and their families.”[456] The Committee has stressed that the minimum wage should “apply systematically” and protect as much as possible “the fullest range of workers, including workers in vulnerable situations.”[457]

For workers with “precarious contracts,” such as app-based workers, the CESCR has emphasized that “supplements to the wage, as well as other measures to guard against arbitrariness, may be necessary in the interest of fairness to mitigate the lack of job security.”[458]

Right to Privacy

The opaque algorithmic models that app-based companies use to assess and set wages and manage workers’ performance are created and refined using personal sensitive data about workers’ locations, habits, and activities. This collection and use of workers’ personal data should respect and protect workers’ internationally recognized right to privacy.

Article 17 of the ICCPR, which the US has ratified, gives everyone the right to “protection of the law” against “arbitrary or unlawful interference with [their] privacy,” and “unlawful attacks on [their] honour and reputation.”[459] The UN Human Rights Committee, the treaty body charged with monitoring implementation of the ICCPR, has found that the collection and storage of personal data by private entities must be regulated by law.[460]

Data protection is an inextricable component of the right to privacy. According to the non-profit group Privacy International, “Data protection is about safeguarding our fundamental right to privacy by regulating the processing of personal data: providing the individual with rights over their data, and setting up systems of accountability and clear obligations for those who control or undertake the processing of the data.”[461] The UN Office of the High Commissioner for Human Rights in 2018 recommended that all States “[a]dopt strong, robust and comprehensive privacy legislation, including on data privacy, that complies with international human rights law in terms of safeguards, oversight and remedies to effectively protect the right to privacy.”[462] The UN Human Rights Committee has similarly recommended that in order to fulfil Article 17 of the ICCPR, “[t]he gathering and holding of personal information ... whether by public authorities or private individuals or bodies, must be regulated by law.”[463] The Committee concluded that Article 17 of the ICCPR also establishes “the right to ascertain in an intelligible form whether, and if so, what personal data is stored in automatic data files, and for what purposes”, and that individuals have the right to “request rectification or elimination” when their personal data is processed incorrectly.

The ILO and various countries have translated these general principles into worker-specific protections. In 1997, the ILO warned that “new ways of collecting and processing data entail some new risks for workers,” and created a Code of Practice to guide the protection of workers’ personal data.[464] Although the code is not legally binding on ILO member states, its recommendations provide a useful framework for understanding and challenging how app-based companies compensate and penalize workers using their personal data. As of 2023, the US was the largest member state and donor of the ILO.[465]

The code specifies that workers have the right to know how their personal data is collected and processed, including “the right to examine and obtain a copy of any records” that include their personal data. The “transparency of the processing and thus the workers’ opportunity to control the use of their data” also requires limiting employers’ access to such data to “purposes known to them [workers] and unequivocally defined before their collection.”[466] Workers should also have the right to “demand that incorrect or incomplete personal data” be “deleted or rectified.”

The United States has not yet developed either a federal data protection law, or meaningful standards to protect workers’ privacy and prevent algorithmic discrimination in the workplace.

Right to Safe and Healthy Working Conditions 

International human rights law protects the right of all people to safe and healthy working conditions. In addition, as reaffirmed by the ILO Convention No. 190 adopted in 2019, Concerning the Elimination of Violence and Harassment in the World of Work, everyone has a right “to a world of work free from violence and harassment.”[467]

The UDHR states that “[e]veryone has the right to … just and favorable conditions of work.”[468] Similarly, the ICESCR guarantees “the right of everyone to the enjoyment of just and favorable conditions of work which ensure … safe and healthy working conditions.”[469] 

International human rights law also recognizes the right of all people to the highest attainable standard of physical and mental health.[470] A component of this right obligates states to take the necessary steps for the “prevention, treatment and control of … occupational and other diseases.”[471] The Committee on Economic, Social and Cultural Rights has affirmed states’ obligations to protect the health of workers. States are obligated under the ICESCR to ensure “[p]reventative measures in respect of occupational accidents and diseases … [and] the minimization, so far as is reasonably practicable, of the causes of health hazards inherent in the working environment.”[472]

The US is not a party to the ILO Convention No. 155, Occupational Safety and Health Convention (1981) or ILO Convention No. 161, Occupational Health Services Convention (1985). However, these ILO conventions are recognized as a statement of international best practices on workers’ right to occupational health. ILO Convention No. 155 calls on states “to prevent accidents and injuries to health arising out of, linked with or occurring in the course of work, by minimizing, so far as is reasonably practicable, the causes of hazards inherent in the working environment.”[473] ILO Convention No. 161 similarly calls on states to identify health hazards in the workplace and to progressively develop occupational health services for all workers, “which will facilitate optimal physical and mental health in relation to work.”[474]

ILO Convention No. 155 also recognizes that “[t]he enforcement of laws and regulations concerning occupational safety and health and the working environment shall be secured by an adequate and appropriate system of inspection [that] shall provide for adequate penalties.”[475]

Right to Social Security

Social security is an internationally recognized human right. Article 22 of the Universal Declaration of Human Rights of 1948, for example, spells out the essential elements of the right:

Everyone, as a member of society, has the right to social security and is entitled to realization, through national effort and international co-operation and in accordance with the organization and resources of each State, of the economic, social and cultural rights indispensable for his dignity and the free development of his personality.[476]

Since then, the right to social security has been guaranteed in countries’ national constitutions and reinforced through a range of other international conventions and frameworks.[477] In its General Comment 19, the UN Committee on Economic, Social and Cultural Rights defines this right to encompass at least nine areas of support: health care, sickness, old age, unemployment, employment injury, family and child support, maternity, disability, and survivors and orphans.[478]

States that are parties to the ICESCR assume the obligations to respect, protect, and fulfill the right to social security along each of these areas of support, including by making these programs available, accessible, acceptable, and adaptable.[479] This also requires providing benefits, whether in cash or in kind, that are adequate in both amount and duration.

More recent international human rights law instruments recognize a right to social protection, in addition to and as distinct from the right to social security. For example, a recent protocol to the African Charter on Human and Peoples’ Rights guarantees both rights separately. According to the Protocol, social security protects against income insecurity caused by events such as unemployment, sickness, or maternity.[480] Social protection encompasses all forms of social security while also including strategies and programs that help ensure a minimum standard of livelihood, health services, and care.

The ILO, which has provided detailed guidance on implementation of the right to social security, has set out a list of “social protection floors”—such as access to health care, child benefits, unemployment benefits, and pensions—which can be considered as reflecting the minimum core contents of the right to social security, and other ILO social security instruments set minimum benchmarks for the protection of economically active persons, including platform workers.[481]

Right to Health

Health is a human right indispensable to the exercise of other human rights, and every human being is entitled to the enjoyment of the highest attainable standard of health conducive to living a life in dignity.[482]

The UDHR recognizes that all people have “the right to a standard of living adequate for the health and well-being of himself and of his family, including food, clothing, housing and medical care and necessary social services, and the right to security in the event of unemployment, sickness, disability, widowhood, old age or other lack of livelihood in circumstances beyond his control.”[483]

Similarly, the ICESCR guarantees the rights to “an adequate standard of living” and “the enjoyment of the highest attainable standard of physical and mental health,” including the “prevention, treatment and control of epidemic, endemic, occupational and other diseases” and the “creation of conditions which would assure to all medical service and medical attention in the event of sickness.”[484]

Governments should ensure that everyone has access to the highest possible standard of health care, regardless of their ability to pay.[485]

Right to Unionize

A widely accepted body of international norms has set forth standards for workers' freedom of association. The UDHR sets out that “[e]veryone has the right to form and to join trade unions for the protection of his interests.”[486] The ICESCR similarly recognizes “[t]he right of everyone to form trade unions and join the trade union of his choice.”[487]

This is further articulated in the ICCPR, which states that “everyone shall have the right to freedom of association with others, including the right to form and join trade unions for the protection of his interests.”[488] As a party to the ICCPR, the US has made a commitment to “take the necessary steps … to adopt such legislative or other measures as may be necessary to give effect to” the right to form and join trade unions and to ensure that any person whose right to organize is violated “shall have an effective remedy.”[489]

Workers' freedom of association in human rights instruments has been complemented by legal guidelines on international labor norms developed in detail by the ILO. These norms set forth the right to organize, the right to bargain collectively, and the right to strike as fundamental rights. They are inextricably tied to the exercise of the right to freedom of association and must be protected by national governments. Nearly every country is a member of the ILO. Each is bound by ILO Conventions 87 and 98 dealing with freedom of association whether or not they have ratified those conventions, since freedom of association is taken to be a constitutional norm binding on countries by virtue of their membership in the organization.[490]

Under ILO Convention 87 “[w]orkers … without distinction whatsoever, shall have the right to establish and, subject only to the rules of the organisation concerned, to join organisations of their own choosing.”[491] This includes self-employed and platform workers.[492]

ILO Convention 98 further explains that all workers should enjoy protection against discrimination (including victimization and retaliation) based on their union activities and employer interference in their organizations.[493]

The US has not ratified Conventions 87 and 98 but has long acknowledged its obligations under them. In 1998, the US championed adoption at the ILO of a landmark Declaration on Fundamental Principles and Rights at Work stating that:

…all Members, even if they have not ratified the Conventions in question, have an obligation, arising from the very fact of membership in the Organization, to respect, to promote and to realize, in good faith and in accordance with the Constitution [of the ILO], the principles concerning the fundamental rights which are the subject of those Conventions, namely: (a) freedom of association and the effective recognition of the right to collective bargaining;…[494]

Corporate Human Rights Responsibilities

Even in the absence of regulations, under the UN Guiding Principles on Business and Human Rights (UN Guiding Principles), all businesses have a responsibility to respect human rights and ensure that they do not cause or contribute to actual or potential adverse human rights impacts.[495]

This refers to internationally recognized human rights; understood, at a minimum, as those expressed in the International Bill of Human Rights and the principles concerning fundamental rights set out in the ILO’s Declaration on Fundamental Principles and Rights at Work .[496] The responsibility of business enterprises to respect human rights includes platform companies as it applies to all enterprises regardless of their size, sector, operational context, ownership and structure.[497]

Fundamental to identifying, preventing, and mitigating adverse human rights impacts is the requirement that companies carry out human rights due diligence to identify the possible and actual human rights impacts of their operations and to establish meaningful processes to prevent and mitigate those risks and remedy adverse impacts.[498]

In particular, the UN Guiding Principles state that businesses’ human rights due diligence “should cover adverse human rights impacts that the business enterprise may cause or contribute to through its own activities, or which may be directly linked to its operations, products or services by its business relationships.”[499] 

Acknowledgments

This report would not have been possible without the people who shared their experiences working for labor platforms in Texas and across the US. Human Rights Watch is deeply grateful for their stories and insight.

This report was researched and written by Lena Simet, senior researcher in the Economic Justice and Rights Division at Human Rights Watch, and Amos Toh, former senior researcher in the Technology and Human Rights Division. Anna Bacciarelli, senior researcher in the Technology, Rights and Investigations Division, provided research and editorial guidance and contributed to writing.

Additional research and editorial assistance was provided by Jack Spehn, senior associate in the Economic Justice and Rights Division, and Ekin Ürgen, associate in the Technology, Rights and Investigations Division, who also prepared this report for publication. Jack Spehn and Yulia Lapina, former senior coordinator with the Development Operations team, provided support validating responses to the survey conducted in Texas.

The report was edited by Arvind Ganesan and Sylvain Aubry, director and deputy director of the Economic Justice and Rights division, respectively; Maria McFarland Sánchez-Moreno, former acting deputy program director; Brian Root, senior quantitative analyst; and Deborah Brown, deputy program director, Technology, Rights and Investigations Division. Specialist Review was provided by the following Human Rights Watch staff: Rachel LaFortune, researcher in the Women’s Rights Division; Antonia Juhasz, former senior researcher, Environment and Human Rights Division; Friederike Mager, coordinator, Advocacy; Carlos Rios-Espinosa, associate director, Disability Rights Division; and Hélène de Rengerve, senior advocate, Advocacy. Michael Bochenek, senior legal advisor, Legal and Policy, provided legal review. External legal review was provided by Elizabeth Wang.

Lance Compa, Senior Lecturer at Cornell University's School of Industrial and Labor Relations, graciously reviewed a draft of the report.

The Rideshare Drivers Union (RDU), especially RDU’s President Nicole Moore, provided invaluable support and collaboration throughout the research, particularly in survey design and outreach to workers across Texas and the country. Their dedication to advocating for platform workers’ rights and sharing insights from the lived experiences of platform work has been instrumental to this report.