From “pink eyed terminators” to a clear-eyed response? UK policy on autonomous weapons

By Richard Moyes

In this new paper, Article 36 analyses UK contributions to the discussion on ‘autonomy’ in weapons systems at the Convention on Certain Conventional Weapons. The UK’s political position that existing law is sufficient and no new regulation in this area is necessary obscures a policy orientation dispersed through its statements that has a strong emphasis and useful detail on human control over systems that apply force on the basis of sensors and machine calculations. Brought together, this could provide the building blocks of an approach around which many actors in the international debate could coalesce.

Despite rejecting the idea of developing a legal response, the UK has noted that there may be merits in a ‘code of conduct,’ and that further efforts should be dedicated to ‘operationalising’ the CCW’s Guiding Principles. In 2020, the UK should focus on consolidating its approach to the subject matter in one place. If the CCW’s 2019 ‘Guiding Principles’ are to provide the basis for structuring further work within the CCW’s meetings then it is principle ‘c’ on ‘human-machine interaction’ that should be given the greatest attention. Bringing together under this principle would be a significant contribution to the multilateral debate – a contribution which would be a modest but important step towards leadership.

For the most part, the UK presents an orientation to the subject matter that has conceptual clarity and accords with the key points of Article 36’s thinking. The UK’s position is characterised by:

a strong emphasis on human control;
an assertion that human control can be exerted through constraints on how systems function (through the target profiles that are used), and on duration and location of that functioning, and on the number of applications of force a system can undertake;
an assertion that both system design and operator training need to contribute to systems being ‘explicable’ or understandable at a functional level;
and these factors and considerations are framed within a recognition that different contexts may require (or enable) different constraints in order for the human control to be sufficient.

Although the use of legal language in certain places risks causing confusion (or reveals certain problems of approach, depending on interpretation) the overall framing of content here is very positive. Yet this constructive content is consistently packaged with politically targeted messaging asserting that existing law is straightforwardly sufficient and that states should resist calls to move towards the negotiation of any further regulation. That politically orientated messaging does much to obscure the content – making it difficult for other states (or other actors in the debate) to take a positive lead from the content being offered.

The UK repeatedly asserts that existing law is wholly sufficient, that no new law is needed, that disagreement in the current debate also shows that no new law is possible, and if it were possible (i.e. through a format that did not have the participation of certain militarised states) then it would be unwelcome.

The problem for maintaining this line is that disagreement amongst different actors also means that there is no clarity of shared position regarding existing law in relation to autonomy in weapons systems. Thus, whilst the UK could argue that the duration and space over which a system can function must be set or defined by a human commander so that they can make judgements about that use in relation to the rules of law, other states may not agree. Similarly, the UK’s contention that a system must be practically understandable to an operator in order for that system to be acceptable under the law is not necessarily a position that all states share. Furthermore, for some actors in the debate there are moral or ethical considerations, or considerations of what is socially prudent, that do not find expression in the current legal framework. The lack of agreement – the divergent views of the subject matter, the differing interpretations of existing law, and calls for new additional rules – all stand as evidence against the claim that existing law is adequate in this area.