Clifford Chance

The UK's AI Opportunities Action Plan, unveiled on 13 January 2025, sets out a bold strategy for harnessing AI to help meet the UK's goals for sustained economic growth. The UK government has endorsed all 50 of its recommendations, with the majority of the immediate next steps scheduled for delivery within the next 12 months.

Tech entrepreneur Matt Clifford was commissioned to produce the action plan, which urges Britain to help shape the AI revolution, stressing the need to be an "AI maker, not just an AI taker". The UK's global leadership on AI safety and its regulatory approach to AI are highlighted as key strengths that the UK should leverage in order to develop a thriving domestic AI ecosystem and widespread use of AI.

"Many will welcome this 'boosterism' from the government in an area where the UK excels", said Jonathan Kewley, co-chair of the Clifford Chance Global Tech Group in a recent BBC news interview. "This is the first post-Brexit push for innovation and growth which looks credible. It is a real opportunity for the UK to go its own way in striking the right balance and partnership between innovation and safety."

AI Opportunities Action Plan Overview

The action plan's recommendations fall into three broad themes, each aimed at achieving its stated goals of leveraging AI to drive economic growth, benefit public services and increase personal opportunities.

1. Laying the foundations to enable AI. Recommendations focus on investing in world-class computing and data infrastructure, ensuring access to talent, and "enabling safe and trusted AI development and adoption through regulation, safety and assurance".

2. Changing lives by embracing AI. This outlines recommendations for leveraging AI to help achieve the missions set out in the government's plan for change, including proposals for developing the UK's AI ecosystem through a push for AI adoption in the public sector. The action plan envisages the public sector shaping the AI market in its role as a large customer, and also anticipates benefits to public services arising from the adoption of high-performing, trustworthy AI at scale.

3. Secure our future with homegrown AI. Here, recommendations build on the previous two sections of the action plan, with the aim of having national champions at the frontier of economically and strategically important AI capabilities by the end of the decade. The government has agreed to establish and empower a new unit to " partner with the private sector to deliver the clear mandate of maximising the UK’s stake in frontier AI".

Investing in digital infrastructure

As part of laying the foundations to enable AI, the action plan stresses the importance of securing a sufficient supply of computational power (often called "compute"). In this regard, the action plan recommends long term planning and investment for the UK's AI infrastructure needs and proposes expansion of the UK's continuing investment in data centres by specifically recommending the development of AI Growth Zones to accelerate build out of data centres.

Matt Taylor, Clifford Chance real estate partner, comments: "As the action plan notes, private investment must play a critical role in providing the underlying "compute" needed to harness and drive the benefits from AI.  Data centre investors, operators and customers will welcome the call to streamline planning approvals and accelerate the provision of clean power alongside other measures aimed at accelerating the buildout of data centres.  The market for investment in development of data centres is buoyant but currently being held back by lack of power and an inconsistent approach to planning, and resolving these issues should ensure that the UK retains its place as a leading market of choice for the data centre sector".

The UK's approach to regulating AI

Current approach

A range of UK legislation and regulation applies to AI, including laws on data protection, cybersecurity, product safety, employment and equality, consumer protection, competition and online safety and sector-specific regulations for areas such as financial services and healthcare. Unlike the EU, which published its AI Act last summer, the UK does not currently have a single, overarching legal framework for AI.

The UK's approach to AI regulation has, to date, focused on context-specific, risk-based and proportionate regulation of AI through the application of existing legal frameworks by regulators. These regulators  were tasked, under the UK's AI White Paper of 2022, with applying cross-sectoral principles such as safety, fairness and transparency. Asked to identify, assess, prioritise and contextualise the specific risks addressed by these principles, many UK regulators (including the FCA, PRA, ICO, CMA and Ofcom) published plans, statements, frameworks and guidance regarding their plans for overseeing AI activities falling within their respective remits.

AI-specific legislation?

The action plan recommends preserving the UK's current pro-innovation approach to AI regulation, considering this "proportionate, flexible regulatory approach" to be a source of strength relative to other, more regulated jurisdictions. Acknowledging that well designed and implemented regulation can fuel safe development and adoption of AI, and that clear rules can give businesses the confidence to innovate and invest, the action plan also cautions against ineffective regulation that could hold back adoption in crucial sectors such as the medical sector.

Last year, the government had signalled a plan to introduce UK AI legislation focused on the most powerful AI models, and the action plan acknowledges a need to act quickly to provide clarity on how frontier models will be regulated. However, the plan is agnostic as to whether such any such regulation should be introduced, focusing instead on continued investment in the AI Safety Institute, the role of regulators and regulatory sandboxes, and AI assurance tools.

The government's response states that it will set out its approach on AI regulation and "consult on proposed legislation to provide regulatory certainty to help kickstart growth and protect UK citizens and assets from the critical risks associated with the next generation of the most powerful AI models".

Role of UK regulators

The action plan redoubles the UK's emphasis on existing regulators as overseers of AI development and use, but seeks a potentially controversial shift in the role of UK regulators. 

The action plan makes recommendations to fund a scale-up in the AI capabilities of UK regulators, and further recommendations aimed at incentivising regulators to enable safe AI deployment in the organisations they regulate. In its response, the government agrees to work with certain regulators to identify their future AI capability needs and how they intend to mitigate AI risks and drive growth, including identifying pro-innovation initiatives such as regulatory sandboxes for AI products in areas with "regulatory challenges but high-growth potential, such as products which integrate AI into the physical world like autonomous vehicles, drones and robotics".  Notably, new requirements are envisaged for regulators with significant AI activity to publish annual reports publicly on how they have "enabled innovation and growth driven by AI in their sector".

While UK regulators are already expected to exercise their protective and supervisory functions in a proportionate manner to avoid unnecessarily curbing innovation, commentators are noting that expectations on regulators to promote AI development and deployment can create tensions with their wider mandates and responsibilities. 

Kate Scott, Clifford Chance technology litigation partner, observes: "The UK approach continues to prefer regulatory responsibility for AI to be devolved to sectoral regulators and – for the moment – not to mirror the EU approach of central, cross-sectoral AI legislation focussed on "high risk" AI systems. The key UK regulators – FCA, PRA, CMA, ICO – have all detailed their expectations with respect to AI compliance, and those same regulators will need to ensure that compliance through enforcement, but without stifling innovation. But I wouldn't expect regulators to shy away from AI investigations – a good example being that, a day after the launch of the AI opportunities action plan, on 14 January 2025, the CMA announced an investigation into Google's search and advertising services, including Google's AI services and whether competitors are prevented from entering the market."

Data access

Several of the recommendations in the action plan relate to "unlocking data assets" in both the public and private sectors. These include:

• making available high impact public data sets to AI researchers and developers;
• building public sector data collection infrastructure and financing the creation of new high-value datasets that meet public sector, academia and startup needs;
• offering access to proprietary data sets alongside compute allocation in enticing researchers and start-ups to establish in the UK; 
•  incentivising researchers and industry to curate and unlock private datasets.

In the EU, the Data Governance Act, the Data Act and various "data spaces" initiatives are creating frameworks and platforms for increased data availability and re-use. The action plan does not make recommendations for similar regulatory structures in the UK, but the UK's Data (Use and Access) Bill lays some initial groundwork for the creation of frameworks for the establishment of 'smart data schemes' potentially in the vein of open banking (see our article on the DUA Bill).

Intellectual property

A key issue in relation to access to data sets for AI development is navigation of intellectual property rights. There has already been a significant amount of copyright infringement litigation relating to AI training; most of the cases have been commenced in the US so far, but the UK courts have also been active.  The action plan recommends that the government "reform the UK text and data mining regime so that it is at least as competitive as the EU", stating that uncertainty around intellectual property is hindering innovation and undermining the UK’s broader AI ambitions and the growth of the creative industries.

In December 2024, the government opened a consultation on AI and copyright which outlined various options for reform, including a potential change to the existing data mining exception in UK copyright law to facilitate the training of AI models using copyrighted material.

"The UK has world-leading content industries who want to ensure that, if their work is used to train AI models, they control the terms of any licensing and receive compensation for the use of their work", explains Don McCombie, Clifford Chance IP partner. "However, the sheer volume of material needed to train a Large Language Model has made it difficult for AI model developers to obtain licences from every relevant copyright owner. The UK government will also be conscious that the copyright laws of many other countries are already more favourable for AI companies than those in the UK. Finding a compromise that satisfies all stakeholders will be difficult for the UK government - I don't envy their task!"

Contracting for AI

The action plan urges the public sector to rapidly pilot and scale AI products and services, and to encourage the private sector to do the same. The action plan's aim is for government purchasing power to help shape new markets in AI and boost the domestic ecosystem, with the plan also envisaging this leading to improvements and efficiencies in the delivery of public services. 

As part of a suite of recommendations relating to launching pilots for AI use, the plan outlines a need for "consistent use of a framework for how to source AI – whether to build in-house, buy, or run innovation challenges – that evolves over time, given data, capability, industry context and evaluation of what's worked".  In its response, the government commits to developing a framework for sourcing AI by summer 2025.  Whether the action plan's desire for "light touch procurement" can be realised remains to be seen, given public procurement rules aimed at maintaining transparency and fair competition and the complexities that can arise in contracting for AI.

"The market dynamics for AI procurement will make it interesting to see how AI providers engage with government procurement," says Zayed Al Jamil, Clifford Chance complex commercial transactions and outsourcing partner. "Big AI players enjoy a lot of negotiating power, and other AI service providers may also have to flow down terms from large developers, which can place constraints on negotiations. Those in the private sector will be watching with interest to see whether any large scale government procurement impacts emerging market norms for AI contracts."

A significant level of investment may be needed for truly effective deployment of AI in the public sector action, including investment in public sector digital skills. The action plan acknowledges this in some areas, for example by including a recommendation for supporting the hire of external AI talent to create a "technical senior civil servant stream".

"AI adoption is not just a technical challenge, it is an organisational challenge," notes Zayed. "Successful implementation hinges on organisational and cultural change as much as procurement. This will need to play a significant role in any push towards widespread public sector use of AI."

For more information on key considerations when procuring AI, see our webinar recording: Contracting for AI.

Next steps

There will be many developments and deliverables to track over the next 12 months, including:

• The Department of Science, Innovation and Technology's (DSIT) long-term compute strategy, including plans for allocating compute and addressing sustainability and security challenges, and approach to international collaboration – to be published in Spring 2025
• First AI Growth Zone at Culham to be delivered and process for selecting further AI Growth Zones to be set out – Spring 2025
• Publication of further details on the National Data Library and data access policy – Summer 2025
• Regulatory Innovation Office update on the identification of priority sectors with high-growth potential and work with relevant regulators to identify pro-innovation initiatives – Spring 2025
• DSIT to appoint AI lead for each government mission, build a cross—government technical horizon scanning and market intelligence capability, and explore partnerships with AI vendors and start-ups – update in Autumn 2025
• DSIT framework for sourcing AI – Summer 2025
• DSIT's AI Knowledge Hub pilot – Summer 2025
• DSIT to work with devolved and local government to identify AI adoption development opportunities to drive growth and, where an opportunity is identified, to incorporate AI adoption objectives into Local Growth Plans within the next twelve months.

"The timescales set out in the government's response indicate that the action plan's recommendations are being treated with urgency and we can expect significant levels of activity over the next 12 months," comments Phillip Souta, Global Director of Tech Policy at Clifford Chance. "Organisations should consider how they may wish to engage with the UK government and regulators on the development of AI policy, plans for unlocking data access, or as part of regulatory sandboxes."