My Lords, the noble Lord, Lord Coaker, is right: we have kept until the end of the day—unfortunately when few people are around—one of the best debates we have had during this stage of the Bill. I thank the noble Lords, Lord Browne and Lord Clement-Jones, and the noble and gallant Lord, Lord Craig, for tabling this amendment. I know that their interest is informed and determined, and I can tell them that it is welcome. Having debated this issue with them now on several occasions, I understand the depth of their concern in this important area. I am grateful to them for the way they have engaged with me and officials and I look forward to further engagement, for we will surely debate these issues in this House for many years to come. I say to the noble Lord, Lord Coaker, that any Government would expect to be accountable to Parliament in respect of matters of such significance.
As with so many issues relating to the rapid march of new technology, this is both complex and pressing. The Government continue to welcome the challenge and scrutiny being brought to this question, and, as I noted on previous engagements, I do not dispute the noble Lords’ analysis of the importance of proper legal consideration of novel technologies. Indeed, I attempted to access the podcast to which the noble Lord, Lord Browne, referred. I do not know whether the Chamber will be delighted or disappointed to learn that, such is the security of my MoD computer, I could not get anywhere near it, so I have still to enjoy the benefit of listening to that podcast, which I intend to do.
As I said, I know that the amendment is extremely well intended and timely, but I hope to persuade your Lordships that the proposed review is not the right means of addressing these issues. However, I assure
your Lordships that the department is alert to these questions and has been working extensively on them over the course of the last 18 months. Indeed, the noble Lords, Lord Browne and Lord Clement-Jones, have been engaging with officials in the department. They might have a better understanding than most of what is taking place.
Setting a requirement for a review in law would actually risk slowing down the work needed to develop the policy, frameworks and processes needed to operate AI-enabled systems responsibly, and to address the legal risks that service personnel might otherwise face. That is an issue of profound importance and one in which the noble and gallant Lord, Lord Craig of Radley, is rightly interested.
Noble Lords will understand that I cannot set out details of the department’s position until these have been finalised, but I can assure your Lordships that work to set a clear direction of travel for defence AI, underpinned by proper policy and governance frameworks, has reached an advanced stage. The noble Lord, Lord Browne, will I am sure have a sense of where that is headed. Key to it is the defence AI strategy, which we hope to publish in early course, along with details of the approaches we will use when adopting and using AI.
These commitments, which are included in the National AI Strategy, reflect the Government’s broader commitment that the public sector should set an example through how it governs its own use of the technology. Taken together, we intend that these various publications will give a much clearer picture than is currently available, because we recognise that these are vital issues that attract a great deal of interest and we need to be as transparent and engaged as possible. I wish specifically to reassure the noble Lord, Lord Coaker, about that.
I know from their contributions, to which I listened, that noble Lords will understand that this AI strategy cannot be the last word on the subject, but I hope that, when we do publish details, your Lordships will be substantially reassured that we are on the right track, and that substantial effort and engagement will follow. There is no end to the march of technology—that is one of the reasons why we have questioned the utility of a snapshot review process—nor will there be an end to our challenge of ensuring that we do the right thing with that technology, especially where grave matters of life and death and national security are concerned.
As we undertake this work, one of our top priorities must be to develop the terminology and vocabulary necessary to ensure we illuminate, clarify and improve understanding and awareness, and to find the right way to debate these issues. This is by no means a comment on any of the discussions that we have engaged on in this House; it is more a general observation on the difficulty of debating concepts such as lethal autonomous weapon systems when there is no definition and different views are not always clearly differentiated.
Are we concerned that AI could usher in a new era of weapons which, whether controlled by a human or not, could result in devastation and atrocities? Or are we concerned at the ethical implications of a machine, rather than a human, taking decisions which result in
the death of even a single human? The answer is both, but the discussion is not best served when it jumps between such disparate topics.
The MoD has to keep pace with the threats that confront this country and consider how to deal with them. When I spoke in Grand Committee, I commented, in response to the noble Baroness, Lady Smith, that context-appropriate human involvement could mean some form of real-time human supervision, which might be called “human in the loop”, or control exercised through the setting of a system’s operational parameters. The noble Lord, Lord Browne, correctly observed that some might call the latter a fully autonomous weapon. But I wonder whether they would use that term, or perhaps more importantly be concerned, if the use case they had in mind was a system mounted on a Royal Navy vessel to defend against hypersonic threats. Such a system might well be lethal—that is, capable of taking human life—but in many ways it would not be considered fully autonomous, even if it detected the threat and opened fire faster than a human could react.
We must be careful to avoid generalisations in this debate. We in the Ministry of Defence have a responsibility to ensure that our position is properly communicated. That is a responsibility we acknowledge, and I say again to the noble Lord, Lord Coaker, that it is a responsibility of which we are cognisant and about which we will be vigilant.
The crucial point, which is also the reason why this amendment is unnecessary, is that all new military capabilities are subject to a rigorous review process for compliance with international humanitarian law. Any determination as to the exercising of context-appropriate human involvement will similarly be done carefully on a specific case-by-case basis. We also adjust our operating procedures to ensure that we stay within the boundaries of the law that applies at the time.
International and domestic frameworks provide the same level of protection around the use of novel technologies as for conventional systems because their general principle is to focus on the action, rather than the tool. These frameworks therefore offer appropriate levels of protection for our personnel. We are committed to ensuring that our Armed Forces personnel have the best possible care and protection, including protection against spurious legal challenges. I think I said in Committee that, earlier this year, we acted to bolster this protection in historical cases through the overseas operations Act.
This is a fascinating and complex area. I hope my remarks provide reassurance to your Lordships that the Ministry of Defence takes these matters very seriously, is already doing all that needs to be done and is planning to be proactive in communicating its approach appropriately to Parliament and the public. On this basis, I suggest that this amendment is not needed. The noble Lord, Lord Browne, has been kind enough to indicate that he will not press it, but I hope that he and other Members of this House will remain engaged with us in the MoD, as we will remain engaged with our international partners and allies, and our own public and civil society, so that we can make rapid progress on these important and challenging questions.