Recently the House of Lords Debated the Report of the AI in Weapon Systems Committee Proceed with Caution.
This is an edited version of what I said
Autonomous weapon systems present some of the most emotive and high-risk challenges posed by AI. We have heard a very interesting rehearsal of some of the issues surrounding use and possible benefits, but particularly the risks. I believe that the increasing use of drones in particular, potentially autonomously, in conflicts such as Libya, Syria and Ukraine and now by Iran and Israel, together with AI targeting systems such as Lavender, highlights the urgency of addressing the governance of weapon systems.
The implications of autonomous weapons systems—AWS—are far-reaching. There are serious risks to consider, such as escalation and proliferation of conflict, accountability and lack of accountability for actions,
and cybersecurity vulnerabilities. There is the lack of empathy and kindness qualities that humans are capable of in making military decisions. There is misinformation and disinformation, which is a new kind of warfare.
Professor Stuart Russell, in his Reith lecture on this subject in 2021, painted a stark picture of the risks posed by scalable autonomous weapons capable of destruction on a mass scale. This chilling scenario underlines the urgency with which we must approach the regulation of AWS. The UK military sees AI as a priority for the future, with plans to integrate “boots and bots” to quote a senior military officer.
The UK integrated review of 2021 made lofty commitments to ethical AI development. Despite this and the near global consensus on the need to regulate AWS, the UK has not yet endorsed limitations on their use. The UK’s defence AI strategy and its associated policy statement, Ambitious, Safe, Responsible, acknowledged the line that should not be crossed regarding machines making combat decisions but lacked detail on where this line is drawn, raising ethical, legal and indeed moral concerns.
As we explored this complex landscape as a committee—and it was quite a journey for many of us—we found that, while the term AWS is frequently used, its definition is elusive. The inconsistency in how we define and understand AWS has significant implications for the development and governance of these technologies. However, the committee demonstrated that a working definition is possible, distinguishing between fully and partially autonomous systems. This is clearly still resisted by the Government, as their response has shown.
The current lack of definition allows for the assertion that the UK neither possesses nor intends to develop fully autonomous systems, but the deployment of autonomous systems raises questions about accountability, especially in relation to international humanitarian law. The Government emphasise the sufficiency of existing international humanitarian law while a human element in weapon deployment is retained. The Government have consistently stated that UK forces do not use systems that deploy lethal force without human involvement, and I welcome that.
Despite the UK’s reluctance to limit AWS, the UN and other states advocate for specific regulation. The UN Secretary-General, António Guterres, has called autonomous weapons with life-and-death decision-making powers “politically unacceptable, morally repugnant” and deserving of prohibition, yet an international agreement on limitation remains elusive.
In our view, the rapid development and deployment of AWS necessitates regulatory frameworks that address the myriad of challenges posed by these technologies. The relationship between our own military and the private sector makes it even more important that we address the challenges posed by these technologies and ensure compliance with international law to maintain ethical standards and human oversight. I share the optimism of the noble Lord, Lord Holmes, that this is both possible and necessary.
Human rights organisations have urged the UK to lead in establishing new international law on autonomous weapon systems to address the current deadlock in conventional weapons conventions, and we should do so. There is a clear need for the UK to play an active role in shaping the nature of future military engagement.
A historic moment arrived last November with the UN’s first resolution on autonomous weapons, affirming the application of international law to these systems and setting the stage for further discussion at the UN General Assembly. The UK showed support for the UN resolution that begins consultations on these systems, which I very much welcome. The Government have committed also to explicitly ensure human control at all stages of an AWS’s life cycle. It is essential to have human control over the deployment of the system, to ensure both human moral agency and compliance with international humanitarian law.
However, the Government still have a number of questions to answer. Will they respond positively to the call by the UN Secretary-General and the International Committee of the Red Cross that a legally binding instrument be negotiated by states by 2026? How do the Government intend to engage at the Austrian Government’s conference “Humanity at the Crossroads”, which is taking place in Vienna at the end of this month? What is the Government’s assessment of the implications of the use of AI targeting systems under international humanitarian law? Can the Government clarify how new international law on AWS would be a threat to our defence interests? What factors are preventing the Government adopting a definition of AWS, as the noble Lord, Lord Lisvane, asked? What steps are being taken to ensure meaningful human involvement throughout the life cycle of AI-enabled military systems? Finally, will the Government continue discussions at the Convention on Certain Conventional Weapons, and continue to build a common understanding of autonomous weapon systems and elements of the constraints that should be placed on them?
The committee rightly warns that time is short for us to tackle the issues surrounding AWS. I hope the Government will pay close and urgent attention to its recommendations.
12th August 2022
Creating the best framework for AI in the UK
30th July 2022
Debate on AI in the UK: No Room For Complacency report
19th December 2021
Launch of AI Landscape Overview: Lord C-J on AI Regulation
2nd October 2021