Fifteen years ago today, Irish Ambassador Daithí Ó Ceallaigh brought down his gavel at a conference in Dublin, signaling the adoption of the Convention on Cluster Munitions. In the years since, 123 countries have joined the treaty, which has proven an effective tool to prevent and remediate the civilian suffering caused by these indiscriminate weapons.
On the same date, five years later, Christof Heyns, then United Nations Special Rapporteur on extrajudicial, summary or arbitrary executions, warned countries at the U.N. Human Rights Council about a new threat to humanity: autonomous weapons systems. His presentation and report kickstarted multilateral discussions about how to address the next generation of warfare. Unfortunately, those talks have stalled, and there is a serious risk that rapid technological developments will outpace diplomatic results.
May 30 marks milestones for both these humanitarian disarmament moments: the 15th anniversary of the cluster munitions treaty, and the 10th of Heyns’ call for a moratorium on development and use of autonomous weapons systems, which would select and engage targets based on sensor processing rather than human inputs.
Although political and procedural hurdles have impeded progress on addressing autonomous weapons systems, proponents of a new treaty should look to the success of the Convention on Cluster Munitions, and the negotiations that led to it, for inspiration.
Convention on Cluster Munitions: An Effective Tool
Cluster munitions are large weapons that contain dozens or hundreds of smaller weapons known as submunitions. They kill and injure civilians at the time of attack, especially when used in populated areas, because the submunitions are dispersed widely. In addition, many of the submunitions do not explode on impact as designed – instead, they linger like landmines, endangering civilians for months or years to come.
The Convention on Cluster Munitions absolutely bans the production, transfer, stockpiling, and use of cluster munitions, as well as assistance with any of those activities. It also requires States parties to destroy stockpiles, clear cluster munition remnants, and assist victims.
As of this May, the Convention has 111 States parties and an additional 12 signatories, and it has done much to achieve its goal of ending the suffering and casualties caused by cluster munitions. According to the Cluster Munition Monitor 2022 report, States parties have destroyed 99 percent their stockpiles, removing 1.5 million cluster munitions and 178 million submunitions from countries’ arsenals. While States parties could do more, they have cleared wide swaths of land and provided a range of aid to victims.
States parties have also complied with the prohibition on use, and the stigma generated by the Convention has pressured some States not party to the treaty to cease their use. The United States, for example, has not used the weapon since 2003, except for one isolated attack in Yemen.
No treaty is a panacea. Russia and Ukraine, neither of which is a party, have both used cluster munitions since Russia’s full-scale invasion of Ukraine. Their use has caused hundreds of casualties for Ukrainian civilians. But these attacks have generated international condemnation and NATO States have rejected Ukraine’s request for transfers of cluster munitions.
Autonomous Weapons Systems: Struggle for New Law
Meanwhile, countries have yet to agree to a comparable treaty on autonomous weapons systems even though 91 countries have called for a legally binding instrument. As with cluster munitions, there are concerns that autonomous weapons systems would be indiscriminate, in this case because they would lack the technology and human judgment to distinguish between legitimate military targets and civilians or surrendering or wounded soldiers. Delegating life-and-death determinations to machines also raises a host of other ethical, legal, accountability, and security issues.
Six months after Heyns presented the issue to the Human Rights Council in 2013, the international debate shifted to the Convention on Conventional Weapons (CCW) because most countries wanted to deal with the issue through an international humanitarian law framework rather than an international human rights law body. Discussions in this forum allowed for countries to refine their policy positions, and convergence has emerged around a “two-tier” approach that would include both prohibitions and regulations on autonomous weapons systems.
Proponents use varied terminology, but the version espoused by many countries as well as Human Rights Watch and the Stop Killer Robots campaign calls for prohibiting autonomous weapons that select and engage targets without meaningful human control and adopting regulations to ensure that all other autonomous weapons systems can only operate with meaningful human control. Civil society organizations, along with the International Committee of the Red Cross and some countries, have also called for prohibiting weapons systems that target humans.
According to this view, meaningful human control requires that systems be understandable and predictable and constrained in the temporal and geographic scope of their operation.
Despite this widespread convergence, the CCW operates by consensus, which has prevented countries from agreeing to negotiate a legally binding instrument since a single country can delay and disrupt the process. Some countries, such as the United States, United Kingdom, Australia, South Korea, and others, prefer voluntary measures. Russia has repeatedly obstructed progress even at that level, seeking to weaken the annual mandate of the CCW Group of Governmental Experts that deals with the issue.
At the recent Experts meeting in Geneva in mid-May, States discussed ways to address what they refer to as “lethal autonomous weapons systems” (LAWS). On the last night, diplomats negotiated until past midnight trying to find compromise on a report of their conclusions, but the language became increasingly watered down. While the draft never included a mandate to negotiate a legally binding instrument, even though that is what most countries wanted, it reflected the convergence around a two-tiered approach and support for requiring elements like predictability and understandability in the use of force.
But by early morning, the only consensus States could reach was language like this weak excerpt from final report of the Group of Governmental Experts:
[T]he Group concluded that:
(a) IHL continues to apply fully to the potential development of LAWS;
Weapons systems based on emerging technology in the area of LAWS must not be used if they are incapable of being used in compliance with IHL.
In essence, it means that it is unlawful to use a weapon system that cannot be used lawfully – a true but rather obvious statement that could have been made before the meeting.
Form over Forum: Lessons from the Cluster Munitions Process
The inability to make progress under the CCW process indicates that it is time to seek another forum, and the Convention on Cluster Munitions offers lessons in that regard. In addition to being a successful humanitarian disarmament instrument, it was achieved through a process that serves as a useful model for autonomous weapons systems treaty negotiations.
Multilateral efforts to address the civilian suffering caused by cluster munitions also began in the CCW, and they were similarly stymied by consensus. In that case, countries pursued an independent process outside of the U.N., known as the Oslo Process, following the precedent of the 1997 Mine Ban Treaty negotiations.
The Oslo Process had four key benefits, which address shortcomings of the CCW forum. While there are some subtle differences, these characteristics can also be true of the U.N. General Assembly, as shown by the negotiations of the Treaty on the Prohibition of Nuclear Weapons. The General Assembly is another possible forum for negotiating an autonomous weapons systems treaty.
First, the Oslo Process was open to everyone, but participants had to commit at the beginning to the same goal, in that case to prohibit cluster munitions and adopt certain remedial measures. That step led to more effective and efficient negotiations by ensuring that participating countries agreed on the underlying purpose from the start. Second, countries also agreed early on to a deadline of less than two years, which expedited the process.
Third, negotiating States were not bound by consensus because their rules of procedure allowed for voting. They still sought to achieve unanimity on all matters and, in the Oslo Process, they never needed to resort to a vote, but the option was there, which prevented the threat of a veto leading to the lowest common denominator.
Finally, the Oslo Process was inclusive, involving a range of countries, international organizations, civil society, and survivors. This characteristic contrasts with the emerging CCW trend, exemplified at the last meeting when Russia insisted that observers be removed from informal sessions over the objections of 21 countries that supported their participation and the principle of inclusion.
Ultimately, countries that are serious about addressing the threats posed by autonomous weapons systems should not let the forum for negotiations dictate the form of their response to those threats. They should not remain beholden to the current CCW process, which seems after 10 years to have run its course. Instead, they should hold firm to their commitment to pursue a treaty on autonomous weapons systems and choose a process that will allow them to achieve it.
On this 15th anniversary, proponents of an autonomous weapons systems treaty can look to the Convention on Cluster Munitions for guidance and motivation. It shows that an effective, efficient, and inclusive process can lead to life-saving results.