US-UK deal envisages policy to support AI adoption

Cerys Wyn Davies, Michael Pocock, and Sarah Cameron of Pinsent Masons were commenting after details of the US-UK ‘technology prosperity deal’ emerged in a ‘memorandum of understanding’ (MoU) published by the two governments. Set out in the document are details of how the US and UK governments intend to collaborate on a number of technological and scientific fronts, including civil nuclear, quantum computing, and AI.

The US and UK described AI as “the defining technology of our age”, adding that it presents “limitless opportunities to improve people’s lives”.

Collaboration on AI policy and adoption is one focus of the new deal. The US and UK also confirmed plans to “collaborate closely in the build-out of powerful AI infrastructure, facilitate research community access to compute, support the creation of new scientific data sets, and harness their expertise in metrology and evaluations to enable adoption and advance our collective security”. They said they intend to “leverage this infrastructure and the AI expertise across industry and elsewhere, to deliver transformational AI-driven change for our societies and economies”.

Joint research initiatives to explore the role AI has to play in enabling science, including in supporting “precision medicine including for cancer and rare and chronic diseases”, are also planned, while the governments said they are intent on “exploring opportunities for collaboration in building secure AI infrastructure and supporting AI hardware innovation” and on “promoting US and UK AI exports to offer the full stack of chips, data centres, and models” too. Further work will focus on the development of AI standards and on ensuring people benefit from the opportunities of AI “across the supply chain”.

Technology law expert Sarah Cameron said: “The challenge for the UK is to maintain a pro innovation approach towards AI – and, in doing so, encourage investment and international partnerships that promote growth – in a way that does not compromise the approach it has committed to around AI regulation with the five core principles of safety, security and robustness; appropriate transparency and explainability; fairness; accountability and governance; and contestability and redress.”

“The US-UK deal aligns with the UK’s AI opportunities action plan’s focus on certain AI ‘foundations’ – like computing and data infrastructure – but there is apparent disconnect in other areas. For example, the UK action plan puts emphasis on ‘enabling safe and trusted AI development and adoption through regulation, safety and assurance’ and specifically states that innovation must come alongside effective assurance and the protection by government of its citizens, particularly the marginalised. Those sentiments are not reflected in the wording of the US-UK MoU, but on the flip side the UK has so far resisted US pressure to remove its digital services tax, which many view as targeted at US-based technology companies,” she said.

Cameron added that the challenge for the UK government is how it stays true to its own commitments while maintaining good relations with important international partners, like the US and EU.

“Much has been made of the UK not signing the declaration from the Paris AI summit earlier this year, and of its renaming of the AI Safety Institute as the AI Security Institute, given the UK championed global leadership on AI safety and governance in the Bletchley declaration from the earlier international AI summit in November 2023,” Cameron said. However, on the ground, since Paris, the AI Security Institute has built an international coalition on safeguarding AI, working with Canada, Amazon, Anthropic and others, that is focused on predictable systems and rooting out harmful behaviours and ensuring transparency, human oversight and alignment with human values.”

“This shows the UK is trying to navigate a careful middle ground on AI in challenging political circumstances,” she said.

Data protection and intellectual property law expert Cerys Wyn Davies said: “As well as staying true to its commitment to the five core principles in addressing AI regulation, the UK should continue to steer a clear path in relation to data protection regulation and copyright.”

“The UK has confirmed its commitment to protect individuals’ privacy rights under the Data Use and Access Act 2025 whilst promoting innovation and economic growth. Following the government’s copyright and AI consultation last winter, the UK is also seeking to achieve an appropriate balance between the interests of content creators and AI developers over use of content to train AI systems, as both groups contribute to innovation and economic growth. It is key that the US-UK deal does not cause the UK to divert from this course. It is therefore encouraging that technology secretary Liz Kendall told the BBC that no guarantees were made to AI companies on copyright, a major issue for the UK’s creative sector,” she said.

Planning expert Michael Pocock said: “The signing of the MoU reinforces the government’s commitment to the wider AI agenda, including data centre development. Over the last 12 months there has been a clear direction of travel in the UK in support of accelerating AI development with a suite of changes to the planning system, including designating AI ‘growth zones’ and changes to the National Planning Policy Framework. The government is also currently considering putting data centres within the ‘nationally significant infrastructure projects’ regime, which could be a game-changer in terms of bringing together the wider infrastructure and associated regulatory consents required by ‘powerful AI infrastructure’ under one umbrella.”

“A common theme for both the UK and the US is the need for neighbouring power generation as part of data centre projects. A collaborative approach to this issue should benefit in terms of sharing further innovation which may assist in finding more efficient and technologically advanced solutions to help remove power constraints,” he added.