Joint committee on Human Rights

Peter Swallow Excerpts

Thursday 15th October 2026 Westminster Hall

[Alex Sobel]

I thank my hon. Friend the Chair of the Science and Technology Committee. It has been my great privilege to twice hear Pope Leo address an audience in the Vatican. The second time, at the Jubilee of Governments, last year in June, he addressed this exact issue. This is not a recent development for the Pope. I read his encyclical and think that every Government should study it. It is interesting that—although he is technically a head of state, because the Vatican is a state—a faith leader such as the Pope has such an incisive view on an issue such as this and brings it back to the central principle that human rights are for humans. We have already seen the anthropocentric effects on the climate. We cannot have the same pervade every area of life through AI. These AI companies are now at a stage where they are trying to replace Governments and not be regulated by Governments. That is, exactly as my hon Friend says, because they can then limit their liabilities, maximise their profits and act without any scope from the state. That should concern us all.

[Peter Swallow (Bracknell) (Lab)]

My hon. Friend will not have failed to clock the comments from the president of Microsoft this week that the UK is best placed to lead on AI governance. That is something that we found in our report, not least because of the AI Security Institute, which our report recommends strengthening further. Does he agree that it is not about limiting the very real benefits that AI may bring to the UK’s economy, our healthcare system or even, as we found in our report, to human rights, but about leading the world in realising those benefits safely while recognising and mitigating the very real risks that exist?

[Alex Sobel]

I absolutely agree. We continually heard evidence right across the industry, both here in the House and externally, about the excellent work that the AISI does. I am pleased that the AI Minister, who is here in the Chamber, has played another leading role in developing AISI as the world’s leading AI safety institute.

However, my hon. Friend is absolutely right. There are two things. The first, as the report found, is about human utilisation of AI tools through, for instance, deepfakes, which affect people’s human rights, or through employment practices, which we can probably very easily regulate, control and legislate for, but that need Government action.

Then, secondly, there are the frontier models. We need a very different level of regulation and legislation for those than we have for individual assistive tools. I am not really worried about ChatGPT 2.0. It is not going to create an existential crisis but, without regulation, future models that are being developed can and will.