II. Article 36 and the Requirement to Review New Weapons
States should review new and modified weapons for their compliance with international law. This rule is codified in Article 36 of Additional Protocol I to the Geneva Conventions, which states:
In the study, development, acquisition or adoption of a new weapon, means or method of war, a High Contracting Party is under an obligation to determine whether its employment would, in some or all circumstances, be prohibited by this Protocol or by any other rule of international law applicable to the High Contracting Party.
Whether considered new types of weapons or modifications of ones that have greater human supervision, autonomous weapons should be subject to such reviews. In fact, the International Committee of the Red Cross (ICRC) specifically highlighted autonomous weapons as an area of concern in its authoritative commentary on Article 36. The ICRC wrote, “The use of long distance, remote control weapons, or weapons connected to sensors positioned in the field, leads to the automation of the battlefield in which the soldier plays an increasingly less important role…. [A]ll predictions agree that if man does not master technology, but allows it to master him, he will be destroyed by technology.” This statement from 1987 raised alarms about the risks of partly autonomous weapons. The warning is even more apt for fully autonomous models.
All states, whether or not they are party to Protocol I, should conduct weapons reviews. Some experts contend that Article 36 is customary international law binding on all states, while others see it as a best practice. The ICRC argues that the obligation applies to all states because “the faithful and responsible application of its international law obligations would require a State to ensure that the new weapons, means and methods of warfare it develops or acquires will not violate these obligations.” Regardless of their opinion of Article 36’s legal status, many weapons producing states have accepted the obligation to review. The list includesthe United States, which is not party to Protocol I but has been a leader in robot research.
The review of weapons, including robotic ones, should take place at the earliest stage possible and continue through any development that proceeds. For a producing state, “reviews should take place at the stage of the conception/design of the weapon, and thereafter at the stages of its technological development (development of prototypes and testing), and in any case before entering into the production contract.” Evaluations of weapons being modified should similarly be started early in the process. States must also review weapons that they plan to acquire rather than produce themselves. Given that certain states, such as the United States, are putting large amounts of money into research and development of autonomous weapons, the time to begin reviews is now. In addition to being required, an early assessment is in a state’s interest because weapons development can be expensive and it makes sense to avoid costs that may produce only an unlawful weapon.
When determining if a review is necessary, states should define “weapon” broadly to encompass major components and final products. There has been some concern that the United States dodged a review of at least one new unmanned system by arguing that evaluations of its individual components were sufficient. The Judge Advocate General’s office reportedly said that it did not need to review a newly weaponized Predator drone because both the Predator, when used for surveillance, and the Hellfire missile with which it was to be armed, had previously passed reviews when considered separately. An ICRC guide to Article 36, however, says reviews should cover “an existing weapon that is modified in a way that alters its function, or a weapon that has already passed a legal review but that is subsequently modified.” This rule is especially important for robots because they are complex systems that often combine a multitude of components that work differently in different combinations.
Reviews should also be sensitive to the fact that some robotic technology, while not inherently harmful, has the potential one day to be weaponized. As soon as such robots are weaponized, states should initiate their regular, rigorous review process. It would be even better to review this technology before weaponization to ensure that robots do not get to that stage, especially since states are more reluctant to give up weapons later in the development process. Such reviews would be designed to preempt fully autonomous weapons that are inconsistent with international humanitarian law, not to block all work in robotics.
The purpose of a weapons review is to determine if the new or modified weapon would be prohibited by international law. First, states should consider prohibitions under existing weapons treaties. While it is possible that fully autonomous weapons could include components banned or regulated by such treaties, there is no existing treaty that prohibits them as a class. States must then evaluate whether a weapon runs counter to other treaties or customary law. Particularly significant for this discussion are the rules of distinction, proportionality, and military necessity, the cornerstones of international humanitarian law, all of which are accepted as customary. The Martens Clause, which prohibits weapons that run counter to the “dictates of public conscience,” may also be relevant.
The requirement of distinction is arguably the bedrock principle of international humanitarian law. According to customary international law, articulated in Protocol I to the Geneva Conventions, combatants must “distinguish between the civilian population and combatants.” Attacks that fail to distinguish are indiscriminate and unlawful. Indiscriminate attacks include those that do not target a specific military objective, “employ a method or means of combat which cannot be directed at a specific military objective,” or “employ a method or means of combat the effects of which cannot be limited.”
International humanitarian law also prohibits disproportionate attacks, in which civilian harm outweighs military benefits. Protocol I defines a disproportionate attack as one that “may be expected to cause incidental loss of civilian life, injury to civilians, damage to civilian objects, or a combination thereof, which would be excessive in relation to the concrete and direct military advantage anticipated.” Determination of proportionality requires a subjective balancing of the military and civilian impacts of an attack as anticipated before it is launched.
Although not clearly articulated in a single treaty, the concept of military necessity “infuses” international humanitarian law. It provides that lethal force may only be used for the explicit purpose of defeating an adversary, it must be proportionate, and it “cannot justify violation of the other rules of [international humanitarian law].” As one scholar described it, “Military necessity dictates that military force should only be used against the enemy to the extent necessary for winning the war.” For example, attacking surrendering or wounded troops would be unlawful because it is not essential for victory and is expressly prohibited by the Geneva Conventions.
Finally, reviews should assess a weapon under the Martens Clause. The clause dates back to the 1899 and 1907 Hague Conventions and was codified more recently in Article 1(2) of Protocol 1, which states:
In cases not covered by this Protocol or by other international agreements, civilians and combatants remain under the protection and authority of the principles of international law derived from established custom, from the principles of humanity and from dictates of public conscience.
In other words, even if a means of war does not violate an existing treaty or customary law, it can still be found unlawful if it contravenes the principles of humanity or the dictates of public conscience. The International Court of Justice, which found the rule to be customary international law, noted that it “had proved to be an effective means of addressing rapid evolution of military technology.” The clause is a useful tool for evaluating and governing emerging weapons because they often develop faster than international law.
States interested in developing or acquiring fully autonomous weapons should initiate detailed legal reviews of any existing or proposed technology that could lead to such robots. These reviews should begin in the early stages of development, address all configurations of the weapons, and consider such key principles of international humanitarian law as distinction, proportionality, and military necessity. They should also take into account the Martens Clause. States should then cease development of any weapons that fail to meet legal requirements before they become so invested in the technology that they will be reluctant to give it up.
 Protocol I, art. 36.
 International Committee of the Red Cross (ICRC), Commentary on the Additional Protocols of 8 June 1977 to the Geneva Conventions of 12 August 1949 (Geneva: Martinus Nijhoff Publishers, 1987), http://www.icrc.org/ihl.nsf/b466ed681ddfcfd241256739003e6368/f095453e41336b76c12563cd00432aa1!OpenDocument (accessed October 30, 2012), pp. 427-428.
 See, for example, Darren Stewart, “New Technology and the Law of Armed Conflict,” in Raul A. “Pete” Pedrozo and Daria P. Wollschlaeger, eds., International Law and the Changing Character of War (Newport, RI: US Naval War College, 2011), p. 283.
 ICRC, A Guide to the Legal Review of New Weapons, Means and Methods of Warfare: Measures to Implement Article 36 of Additional Protocol I of 1977 (Geneva: ICRC, 2006), http://www.icrc.org/eng/assets/files/other/icrc_002_0902.pdf (accessed October 30, 2012), p. 4.
 States known to have formal review mechanisms include Australia, Belgium, the Netherlands, Norway, Sweden, the United Kingdom, and the United States. Ibid., pp. 5-6, n. 8. For the US requirement to review, see US Department of Defense Directive 5000.1: The Defense Acquisition System, Defense Acquisition Guidebook, E1.1.15 (Legal Compliance), May 12, 2003 (certified current as of November 20, 2007), https://acc.dau.mil/CommunityBrowser.aspx?id=314790&lang=en-US (accessed October 30, 2012).
 ICRC, A Guide to the Legal Review of New Weapons, Means and Methods of Warfare, p. 23.
 Ibid., p. 24.
 Noel Sharkey, “Killing Made Easy: From Joysticks to Politics,” in Lin, Abney, and Bekey, eds., Robot Ethics, p. 119 (citing John Canning et al., “A Concept for the Operation of Armed Autonomous Systems on the Battlefield,” paper presented at the Association for Unmanned Vehicles Systems International conference, Anaheim, CA, August 17, 2004).
 ICRC, A Guide to the Legal Review of New Weapons, Means and Methods of Warfare, p. 10.
 Human Rights Watch telephone interview with Noel Sharkey, September 6, 2012.
 ICRC, A Guide to the Legal Review of New Weapons, Means and Methods of Warfare, pp. 11-14.
 For example, a fully autonomous weapon carrying biological or chemical weapons would be illegal under the Biological Weapons Convention or Chemical Weapons Convention.
 ICRC, A Guide to the Legal Review of New Weapons, Means and Methods of Warfare, pp. 16-17.
 Ibid., p. 17.
 Article 48 of Additional Protocol I to the Geneva Conventions codifies this rule. See Protocol I, art. 48. See also ICRC, “Rule 1: The Principle of Distinction between Civilians and Combatants,” Customary International Humanitarian LawDatabase, http://www.icrc.org/customary-ihl/eng/docs/v1_rul_rule1 (accessed October 4, 2012).
 Protocol I, art. 51(4).
 Ibid., art. 51(5)(b).
 Michael Schmitt, “Military Necessity and Humanity in International Humanitarian Law: Preserving the Delicate Balance,” Virginia Journal of International Law, vol. 50 (2010), http://www.usnwc.edu/getattachment/3c6d9d5d-71c9-45b4-b5ad-99526e06849e/Military-Necessity-VJIL-50-4-Schmitt-Essay (accessed October 30, 2012), p. 835.
 Françoise Hampson, “Military Necessity,” Crimes of War (online edition), http://www.crimesofwar.org/a-z-guide/military-necessity/ (accessed October 4, 2012). Military necessity has also been defined as justifying “measures of regulated force not forbidden by international law which are indispensable for security and the prompt submission of the enemy, with the least possible expenditures of economic and human resources.” Richard M. O’Meara, “Contemporary Governance Architecture Regarding Robotics Technologies: An Assessment,” in Lin, Abney, and Bekey, eds., Robot Ethics, p. 161 (quoting Roy Gutman and Doaud Kuttab, “Indiscriminate Attack, Crimes of War 2.0: What the Public Should Know (2007), pp. 239-241).
 Krishnan, Killer Robots, p. 91.
 Geneva Convention Relative to the Protection of Civilian Persons in Time of War (Fourth Geneva Convention), adopted August 12, 1949, 75 U.N.T.S. 287, entered into force October 21, 1950, art. 3(1); Protocol I, art. 41.
 ICRC, A Guide to the Legal Review of New Weapons, Means and Methods of Warfare, p. 17.
 Protocol I, art. 1(2). See also the preambles of the 1899 and 1907 Hague Conventions. Hague Convention with Respect to the Laws and Customs of War on Land and its Annex: Regulation Concerning the Laws and Customs of War on Land (Hague Convention II), adopted July 29, 1899, 32 Stat. 1803, 1 Bevans 247, 26 Martens Nouveau Recueil (ser. 2) 949, 187 Consol. T.S. 429, entered into force September 4, 1900, pmbl.; Hague Convention Respecting the Laws and Customs of War on Land and its Annex: Regulations Concerning the Laws and Customs of War on Land (Hague Convention IV), adopted October 18, 1907, 36 Stat. 2277, 1 Bevans 631, 205 Consol. T.S. 277, 3 Martens Nouveau Recueil (ser. 3) 461, entered into force January 26, 1910, pmbl.
 Legality of the Threat or Use of Nuclear Weapons, International Court of Justice, Advisory Opinion, July 8, 1996, paras. 78, 84.
 According to the ICRC, the Martens Clause “should be seen as a dynamic factor proclaiming the applicability of the principles mentioned regardless of subsequent developments of types of situation or technology.” ICRC, Commentary on the Additional Protocols of 8 June 1977 to the Geneva Conventions of 12 August 1949 , http://www.icrc.org/ihl.nsf/COM/470-750004?OpenDocument (accessed October 30, 2012), p. 39.