Switzerland’s dangerous neutrality on «Killer Robots»

With the rise of Artificial Intelligence (AI), humankind has outsourced many tasks to its robotic counterparts. However, while the inclusion of AI in everyday objects, such as cars and smartphones, has occurred in front of society’s eyes, a much more secretive revolution has proceeded in the laboratories of weapons manufacturers around the world: the development of AI in weaponry. Switzerland, with its strong humanitarian tradition, as the depositary state of the Geneva Conventions and host state of the ICRC must ensure that international law in general and international humanitarian law in particular keep up with the pace of this technology.

The (non-)regulation of LAWS

Many experts have acknowledged the danger that emanates from these so-called «Lethal Autonomous Weapons Systems» (LAWS, commonly also referred to as «killer robots»), which do not yet exist, but could one day choose and eliminate targets without human involvement. That is why multilateral diplomatic efforts are undertaken at the UN in Geneva, in order to determine if and how such weapons can be regulated or even pre-emptively banned. However, the current arms control regime is under serious threat, as many landmark non-proliferation treaties are running out or are being abandoned. Therefore, the outcome of efforts to regulate LAWS, undertaken under the framework of the Convention on Certain Conventional Weapons (CCW), does not look very promising.

Furthermore, there are three major obstacles specific to the regulation of LAWS. First, LAWS could generate an enormous comparative military advantage. Thus, states currently developing this technology do not want to restrain themselves by ratifying a treaty. Second, the lack of an agreed definition of «autonomy» complicates efforts to regulate LAWS by an international agreement. Third, the «weapons of rich nations», like nuclear weapons, tend to be prohibited or regulated much less than the «weapons of poorer nations», like anti-personnel mines. The rich nations that develop LAWS, amongst them the US, China and Russia, are powerful and unlikely to give up their expensive technology voluntarily.

The Swiss position

Switzerland’s position with respect to LAWS has been, as is often the case, one of a middle ground. While at least 28 countrieshave called for an outright and pre-emptive ban on the development and use of LAWS, the Federal Council declared in 2017 that Switzerland would not align itself with this position, deeming such a ban «premature», because states may also end up banning «systems that could be useful, that help prevent collateral damage, for example.» However, while it is true that autonomy in weapon systems could in some instances bring about humanitarian benefits, the dramatic consequences of a LAWS’ malfunction clearly outweigh benefits. Such a possible malfunction was described by computer scientist Noel Sharkey, chairman of the International Committee for Robot Arms Control, in 2008: «Most soldiers would not for example blow up a school full of children if there is a sniper on its roof, but who knows what a robot would do».

Nevertheless, while Switzerland opposes a pre-emptive ban, it acknowledges that autonomy in weapons systems must at least be regulated in some form. In a working paper submitted to the Group of Governmental Experts (GGE) at the CCW, it presented its «compliance-based approach» towards LAWS. In said document, Switzerland reiterates that the development and employment of such weaponry «must remain in compliance with existing international law and, in times of armed conflict, particularly with international humanitarian law (IHL)» (emphasis added). However, IHL, written at a time when AI was mere science fiction, does not cover the eventuality of a «killer robot gone rogue» committing war crimes. Thus, in the case of fullyautonomous weapons systems (i.e. systems where no human can override a decision taken by the machine once it is activated and sent on its mission), potential legal lacunae could pose serious challenges to individual criminal responsibility and state accountability in cases where the rules of armed conflict are infringed upon.

Conclusion

Although Foreign Minister Ignazio Cassis made clear that Switzerland has no intention of developing or acquiring LAWS, that is not enough. Switzerland should advocate for a legally binding instrument that regulates the development, production and use of LAWS, in order to ensure that it will never be possible for machines to have the final say when it comes to decisions on the life and death of human beings. If such a treaty does not ban LAWS entirely, it should at least institute a mechanism ensuring that states respond to any damage caused, no matter the circumstances. Pre-emptive regulation for LAWS is imperative, because once they have been developed, they will eventually be deployed and inevitably malfunction. The ICRC Commentary on article 36 of the Additional Protocol I to the Geneva Conventions forecasts that «if man does not master technology, but allows it to master him, he will be destroyed by technology». In order to avoid this dystopia, the international community must act now.

Image by Lt. Col. Leslie Pratt / Public domain on Wikimedia