Graduating into a global pandemic was never going to be easy.
But for some in the high school class of COVID-19, their academic futures were further jeopardized by a predictive algorithm that assigned them final grades without explanation or means for meaningful redress.
It’s the latest example of how decisions that affect people’s lives are being dangerously made by impenetrable calculations — and it’s one that raises questions about children’s rights under the European Union’s data protection law.
Education systems across Europe struggled this year with how to determine students’ all-important final grades. But one system, the International Baccalaureate (“IB”) — a high school program that is highly regarded by European universities, and offered by both public and private schools in 152 countries — did something unusual.
Having canceled final exams, which make up the majority of an IB student’s grade, the Geneva-based foundation of the same name hastily built an algorithm that used a student’s coursework scores, predicted grades by teachers and their school’s historical IB results to guess what students might have scored if they had taken their exams in a hypothetical, pandemic-free year. The result of the algorithm became the student’s final grade.
The results were catastrophic. Soon after the grades were released, serious mismatches emerged between expected grades based on a student’s prior performance, and those awarded by the algorithm. Because IB students’ university admissions are contingent upon their final grades, the unexpectedly poor grades generated for some resulted in scholarships and admissions offers being revoked.
Principals from Britain, Hong Kong, India, Finland and Norway protested their students’ anomalous grades. Nearly 25,000 parents, students and teachers petitioned the IB to “take a different approach with their grading algorithm and to make it fairer.”
The IB has justified using the algorithm as “the fairest approach we can take for all our students.” But substituting a mathematical approach for human judgment does not automatically make it fair. Algorithms, in fact, are built by people, their outcomes shaped by the lived experiences and beliefs of their creators.
Indeed, other exam boards — including Scotland's and the United Kingdom’s — also came under fire for deploying statistical systems that may be biased against disadvantaged and minority children, prompting Scotland's First Minister Nicola Sturgeon to officially apologize. Scotland backtracked this week, scrapping downgrades for some 75,000 pupils.
Thousands of students were also wrongly and forcibly deported from the U.K. last year based on a flawed algorithmic assessment of their English proficiency exams.
The contention from the IB and other exam boards that a child’s future can be determined from broad historical grading patterns instead of their individual achievements is fundamentally at odds with the premise that education can open doors to opportunity, based on equally recognizing every child’s merit and potential.
Students who have struggled to overcome low expectations and challenging context know that their lives cannot be easily captured, quantified, or even identified by an algorithm. Worse, they would likely be overlooked by an algorithm in favor of the average.
The IB’s algorithm rendered judgment using data that mostly relied on contextual factors outside of students’ own performance. Its reliance on schools’ historical data — using trends of past students to guess how a single student might have achieved — risked being unfair to atypical students, such as high performers in historically low-performing schools.
Because 94 schools new to the IB lacked sufficient data, their students’ grades were adjusted to fit the past performance of students from other schools, who may have had different circumstances. And the IB’s “active” use of teacher-predicted grades was puzzling, absent an explanation of how it would mitigate acknowledged systemic inaccuracies.
To make matters worse, for students looking to contest their results, it is difficult to do so. The IB has refused to explain or reproduce its calculations for any student’s grade, hampering students from effectively contest their assigned grades.
This is fundamentally unjust — and against European law.
When high-stakes decisions — like the grades that can shape a child’s opportunities in life — are calculated through fully automated processes, it’s vital to clearly explain the results to allow students to contest or fix incorrect decisions.
Europe’s General Data Protection Regulation (GDPR) protects this right to an explanation; the Norwegian government has determined that the IB’s process was “unfair” under this law, and is intending to order the IB to redo the “inaccurate” grades.
The IB unquestionably faced an unprecedented situation, and any alternative to award final grades would have been an imperfect assessment. But by narrowly defining its problem as awarding grades as if it were a normal year, it failed to recognize the inherent unfairness of grading students as if their lives had not been upended by a global pandemic.
The IB had alternatives. Instead, it could have used students’ actual academic performance and graded on a generous curve. It could have incorporated practice test grades, third-party moderation to minimize grading bias and teachers’ broad evaluations of student progress.
It could have engaged with universities on flexibly factoring in final grades into this year’s admissions decisions, as universities contemplate opening their now-virtual classes to more students to replace lost revenue.
It increasingly seems like the greatest potential of the power promised by predictive data lies in the realm of misuse.
For this year’s graduating class, who have already responded with grace and resilience in their final year of school, the automating away of their capacity and potential is an unfair and unwanted preview of the world they are graduating into.