Frontier AI Safety

The UK is a founding member of the AI Safety Institute international network. The network convenes for the first time today in San Francisco.

The UK is a global leader in AI development, which brings many opportunities, but we know that the risks associated with AI can be managed only by global co-operation. What steps is the Department taking to ensure that the UK works with other main leaders in AI development, including the US and China, to ensure that the most advanced frontier AI models are safe for global consumers?

My hon. Friend is completely right to say that safety has to be there from the outset. We want our country to safely explore all the opportunities that AI offers, but it can do so only if people are reassured that safety is there from the outset. The UK safety institute is at the forefront of this. It is the first safety institute, and we are at the forefront of delivering international as well as domestic safety. We are currently working on an international review of the science of AI safety, which draws on the expertise of 30 countries.

I call Steve Darling.

On question number 13, Mr Speaker, may I ask the Minister what steps he will take to ensure that people who are visually impaired are able to engage—

Order. Sorry—we have not reached that question. I call the Chair of the Science, Innovation and Technology Committee.

My hon. Friend the Member for Folkestone and Hythe (Tony Vaughan) rightly raises the need for research into frontier AI safety, and I welcome the Government’s commitment to protecting the public from future AI risks. But AI affects all of our lives already. Today, my Committee launches an inquiry into algorithms, AI and their role in spreading online harm, as we saw in the terrible riots over the summer. As we build our evidence, how is the Minister building the evidence base on AI online harms and their social impact right now?

I am extremely grateful to the Chair of the Select Committee for choosing this as her first inquiry. It is an incredibly important area. This Government are committed to the algorithmic transparency recording standard. The previous Government reneged on their commitment to having individual Departments releasing their standard statement each year. This Government are committing to doing so again and will remain committed to reinforcing the fact that algorithms are there to serve people and not the other way round.

There are many firms in Northern Ireland that have the capability and the experience to offer some advice on getting scientific research on AI safety. I know that the Minister is very interested in Northern Ireland, so has he had an opportunity to speak to companies in Northern Ireland so that we can play our part in how we take this matter forward?

I am grateful to the hon. Gentleman for acknowledging my interest in Northern Ireland, which I have already visited since being appointed in order to meet some of the pioneering tech companies there. I will stay committed to ensuring that the Government recognise the talent across Northern Ireland, harnessing it for not just the domestic good but the global good.