AI Systems: Risks

Viscount Camrose Excerpts

Thursday 8th January 2026 Grand Committee

[Viscount Camrose (Con)]

I join noble Lords in thanking my noble friend Lord Fairfax for securing the debate and for speaking so powerfully, as ever, on the subject.

When I attended my first university course on AI in 1999, AGI was more a theoretical thought experiment than a serious possibility, so the warnings from the director-general of MI5 and, as we have heard, from a great many others are not just testament to the dangers of these technologies but equally a reminder that this tech is moving far faster than anyone has expected.

For today, I will confine my brief remarks to what we consider to be three essential requirements if frontier AI is to remain safe and controllable. First, we need a shared definition of artificial general intelligence. At present, Governments, companies and researchers use the term “AGI” to mean a huge variety of different things. Indeed, the very helpful notes from the Library for this debate found it necessary to include a definition and, whether we accept its definition or not, the important thing is that we need to reach a point where we all agree on what we are talking about. I suggest that, without a common, internationally accepted definition, regulation will always lag behind capability.

Secondly, I am afraid that national approaches alone will not work because the AI stack—that is, the hardware, software, data, finance, energy and skills for any significant AI tool—must necessarily be spread across the globe. A number of global and multilateral organisations are working towards global standards for AI safety, but not so far with sufficient co-ordination or impact to provide much reassurance. I suggest, as others have, that the UK has an unrealised opportunity to lead here more.

Thirdly, we all know that we need dynamic alignment of AI with our human societal goals, but we do not know what that means in practice, how it translates to technical and procedural rules and how such rules are to be deployed and enforced. Again, the AISI and other UK bodies are surely well placed to drive that thinking forward globally and I hope the Minister will comment on the UK’s role here.

I close by somewhat unwillingly quoting Vladimir Putin: whoever becomes the leader in this sphere will become the ruler of the world. In other words, for some nations and for some organisations, the incentives to push ahead at speed outweigh the incentives to do so safely. This is a global problem, but we in this country have an opportunity to show leadership. I hope that we take that opportunity.