Skip to main content

AI Safety Research: UK-US Partnership

Volume 748: debated on Tuesday 16 April 2024

I am pleased to inform the House that myself and US Secretary of Commerce Gina Raimondo, on behalf of the UK and the US, signed a memorandum of understanding on 1 April.

The MOU will enable our AI Safety Institutes to work together to follow through on commitments made at the AI Safety Summit held last November at Bletchley Park. It will allow us to develop an interoperable programme of work and approach to AI safety research. This will help us achieve our shared objectives of ensuring the safe development and use of advanced AI.

Specifically, through this MOU, we intend to engage in the following joint activities:

Develop a shared approach to model evaluations, including the underpinning methodologies, infrastructures and processes.

Perform at least one joint testing exercise on a publicly accessible model.

Collaborate on AI safety technical research, to advance international scientific knowledge of frontier AI models and to facilitate sociotechnical policy alignment on AI safety and security.

Explore personnel exchanges between the UK and US AI Safety Institutes.

Share information with one another across the breadth of their activities, in accordance with national laws and regulations, and contracts.

The institutes are already working together to align their scientific approaches, and to accelerate and rapidly iterate robust suites of evaluations for AI models, systems and agents. This will put us in a good position to evaluate the next generation of advanced AI models.

I launched the AI Safety Institute at the AI Safety Summit at Bletchley Park in November last year, making it the first state-backed organisation focused on advanced AI for the public interest. At that time, we set out our ambition for the UK AI Safety Institute to advance the world’s knowledge of AI safety by carefully examining, evaluating, and testing new types of AI. We have delivered on this intention by developing the capabilities and capacity of our institute into a world-leading organisation. The AI Safety Institute is conducting the world’s first Government evaluations of advanced AI systems. We aim to push the frontier by developing state-of-the-art evaluations for safety-relevant capabilities and conducting fundamental AI safety research. The institute will share its progress with the world to facilitate an effective global response to the opportunities and risks of advanced AI.

This formal partnership provides the basis for further international AI safety co-operation. The UK and US AI Safety Institutes will work with other countries to promote AI safety, manage frontier AI risks, and develop linkages between countries on AI safety research. To achieve this, the institutes will work together to develop international standards for AI safety testing and other standards applicable to the development, deployment of use of frontier AI models. We will progress this international collaboration bilaterally and multilaterally in existing multilateral fora, including the upcoming AI Seoul Summit, to be co-hosted by the UK and the Republic of Korea next month.

In closing, I reaffirm this Government’s commitment to tackling the challenges posed by AI head-on. Through collaboration, innovation, and shared determination, we will continue to lead the way in ensuring a safer and more responsible AI landscape for generations to come.