“We reject autonomous weapon systems that are beyond the control of humans. We want to condemn them worldwide.” This concise statement is part of the coalition agreement (para. 7027) of center-right CDU/CSU and center-left SPD. It reflects the general opinion voiced by all parties across the spectrum as well as by researchers and civil society: no machine should “decide” on terminating a human life without human involvement. But the unisono harmony is deceptive as core issues remain disputed. A recent parliamentary hearing by the Bundestag’s Committee on Education, Research and Technology Assessment gives a renewed insight into the positions of the Government, the opposition, and researchers.
Autonomous weapons, or better “autonomy in weapon systems” (also Lethal Autonomous Weapon System or LAWS), is an emerging disruptive technology that is dominating discussions on the war of tomorrow. The exact use, definition, and handling of autonomy in weapons is disputed among researchers and diplomats alike. Nonetheless, since 2014, political talks have been held in the Group of Governmental Experts (GGE) established under the Convention on Certain Conventional Weapons (CCW) to discuss legal implications pertinent to LAWS and to explore common ground for a potential regulation. As a chapeau convention with annexed protocols on specific weapons, the CCW offers an ideal platform for the discussion of such emerging weapon technologies. The member States are, however, divided. On the one side, key military actors reject any multilateral agreement on LAWS at this point. On the other side, proponents of a binding prohibition demand a new CCW protocol, while the middle ground pleas for soft law regulation. Differences exist not only on whether any regulation is necessary at this point, but also on the definition of LAWS (and human involvement) and on what a regulation could look like – including what legal form it should take (find an overview here, p. 17 ff). As the CCW requires consensus amongst its signatories for the adoption of a new protocol, the talks have so far yielded only slim results in form of “possible guiding principles” (see reports 2018, p. 4 and 2019, p. 3). In these negotiations, Germany pushes for a universal soft law regulation and takes on a mediator role between proponents and opponents of an enforceable prohibition. But is this stance broadly supported by national stakeholders? What aspects of LAWS does the German debate focus on?
While there is overall agreement in Germany that some form of multilateral agreement on LAWS is desirable, differences exist on how to define autonomy. The recent parliamentary hearing saw one researcher criticize the definition chosen by the German Ministry of Defense as too narrow (also here, p. 4 and here, p. 127). A similar definition submitted by Germany at the GGE equally leaves some open questions. After all, as the concept of autonomy in weapons is just emerging, the CCW sees all actors condemning LAWS – just each under their own, sometimes unrealistically broad definition. However, as some CCW States now argue that a regulation might be feasible without agreeing on all aspects of a definition (see e.g. Germany), both the international and the German debate see the most disagreement in the third core issue: Which legal instrument should be pursued? This simultaneously raises an old (legal) dispute: What is soft law’s worth?
Both the Green Party and the Left Party have strongly criticized the German soft law approach. At the parliamentary hearing and in three bills dismissed by the Bundestag earlier this year, they have called for a hard law ban on LAWS, if need be outside the CCW, jointly with those States willing to commit themselves (none of which have the intention or capabilities of developing or deploying autonomous weapons). The German left opposition would prefer a binding treaty without key military actors over a continuation of the soft law, global consensus approach. But is their skepticism valid?
The general argument brought against soft law is a lack of compliance due to its non-binding nature. Non-universal hard law on the other hand, is argued to have a pull-effect, eventually forcing the hesitant States to join. In this argument, working towards soft law is prematurely giving up on a binding norm. But a closer look at the reality of arms control paints a different picture. While the tautology of soft law is hardly comprehensible to domestic lawyers, it has become an important part of international governance. The main benefit of resorting to non-binding “norms” is that States are more likely to conclude such agreements. The reluctancy towards a legal commitment can be countered with flexible soft law. This flexibility allows States to later deviate from their agreement if circumstances change (see MPEPIL, para. 6 and here, p. 423). With autonomy in weapons, one of the key obstacles preventing arms control is that States do not want to prematurely limit their military capabilities. Soft law could establish ground rules without restricting the research of potentially beneficial autonomy in weapons. At the same time, if the development later shows that a ban is necessary, early soft law can be a steppingstone on which to build a subsequent hard law treaty. But even without a later treaty, the normative content of early soft law may become part of customary international law down the line. This is supported by examples in international law, most prominently the study on customary international humanitarian law, for which the International Committee of the Red Cross extensively consulted numerous soft law bodies as its sources.
Additionally, compliance in arms control does not seem to be linked to any legal form. Hard law treaties have been violated (see here, p. 60), while soft law – though more difficult to measure – has had positive effects (MPEPIL, para 16, see also here). At the same time, arms control relies on involving key players. A binding treaty banning autonomous weapons without any States in possession of such technology will not affect the deployment of LAWS any more than any other moral, soft obligation on those States. Pressing forward with a non-universal hard law treaty promises little effect on key military States to follow. A prominent example is the Convention on Cluster Munition, which started within the CCW but was moved outside due to a lack of universal support for a ban. While the Convention on Cluster Munition now counts 110 State Parties, the key States manufacturing and using cluster munition have yet to join.
As these examples show, both soft law and hard law instruments can yield positive results in armament control. Soft law has enough of a positive track record to not be disregarded preemptively as an effective tool. In fact, any agreement would beat the current void in regulation. The German Foreign Ministry should thus double down its efforts in pushing to reach the best possible agreement, regardless of the exact legal tool eventually chosen.
Lorenz Rubner is a PhD candidate at the Institute for International Law of Peace and Armed Conflict (IFHV), Ruhr-University Bochum.