Research: ”Ethical but Lethal” – the UK Position on AI Safety and Autonomous Weapons Systems

Naomi Colvin · Programme Director, Blueprint for Free Speech, UK · Bio

Download Full Research Paper (28 November 2024)

This research focuses on the development of autonomous weapons systems (AWS), together with attempts to limit that development and use. It looks at the UK government's position on AWS both domestically and internationally, particularly in relation to the country's broader AI strategy.

We are at a critical moment in the international discussions around the control, prohibition and proliferation of autonomous weapons systems (AWS) and the position of the UK on this issue is important. While the UK was one of 164 countries who signed up to a declaration in the UN General Assembly at the end of 2023, it is also part of the group that has hampered progress in conventional arms control negotiations (https://press.un.org/en/2023/gadis3731.doc.htm).

Internationally, the UK position has been one in favour of voluntary guidelines for the development of autonomous weapons, rather than explicit prohibitions and controls. On a national level too, the UK Government has been reluctant to define exactly where the lines should be drawn. In 2018, the Ministry of Defence (MoD) claimed the UK had no plans to develop “fully autonomous weapons”, but today we know that the UK MoD is prioritising research into the use of AI for military purposes, including AI-enabled autonomous weapons.

On several occasions in the post WWII era, international society has managed to limit the proliferation and use of particular military horrors, perhaps most notably chemical and biological weapons. The 1997 Ottawa Treaty, which prohibited anti-personnel landmines – a basic kind of AWS – has provided an example and inspiration for international civil society’s efforts to Stop Killer Robots.

It is not only campaigners who have become frustrated with the intransigence of the UK Government on this issue. Through 2023, a dedicated House of Lords committee conducted a Special Inquiry into the Use of Artificial Intelligence in Weapons Systems. (https://lordslibrary.parliament.uk/ai-in-weapon-systems-committee-report-proceed-with-caution/)  The Inquiry presented an opportunity to provide some clarity on some fundamental issues where the UK position to date has been purposely obscure – and, perhaps, also at odds with what the British public would want to see their government doing.

The topic of AWS in the UK brings together a number of distinct areas and Naomi Colvin’s fellowship project examines the following issues:

How should we understand and conceptualise AWS? What is the current state of development of autonomous weapons systems in the UK? How far has the UK’s position on LAWs been justified, or even articulated, at the domestic level? Is the UK’s position on AWS compatible with its wider technology policy, which has placed a major emphasis on the safety of the most advanced AI systems? How feasible or adequate are the main proposals for regulation of AWS in light of what we see in conflict zones in 2024? Does the increasing convergence between military and civilian technology suggest alternate strategies for regulation?

The project takes UK policy on AI and international arms control negotiations as its starting point and examine connections and disjuncts between the two, informed by discussions with individuals involved in policy making and the wider AI safety community.