Prepared by Matt Clifford CBE, the Action Plan is a series of 50 recommendations on how to make the most of AI opportunities, not least to drive economic growth and improve public services and interaction with government. The Action Plan focuses on three core goals: “Invest in the foundations of AI”, “Push hard on cross-economy AI adoption” and “Position the UK to be an AI maker, not an AI taker”.
The recommendations, addressing a wide range of actions, have been generally accepted by the Government. Whilst not everything is new, the Government response is a helpful summary of what we can expect to see in the coming months and years. Amongst other things:
- Improved access to data: for example the creation of a National Data Library to allow greater access for AI researchers and innovators to public sector data assets, including at least five “high impact” data sets that will be “unlocked” in a secure, responsible, privacy-preserving and ethical way. We await more detail on the National Data Library and how the Government will manage the disclosure of public sector data sets in Summer 2025. Given the inevitable inclusion of NHS health data in the proposals, there is likely to be significant scrutiny on the safety and security of such disclosure. The plans will outline how the Government intends to take a more strategic approach to shaping the data it collects (rather than simply making available data it already has) and we can also expect more detail regarding the creation of public sector data collection infrastructure. The Action Plan recommendations also call for researchers and industry to be incentivised to make private data sets available.
- Regulation and support for regulators: for example after a significant lead time, a consultation on the Government’s approach to AI regulation, specifically regarding the most powerful frontier models, is promised in Spring 2025. As trailed, the AI Safety Institute will be placed on a statutory footing and more investment is expected to support an emerging AI assurance ecosystem and safety market. The Spring 2025 Spending Review will factor in the need for further funding to upskill and scale regulatory AI capabilities, though previous funding increases have been limited. Where considered necessary and legislation allows, Government departments will be encouraged to promote safe AI innovation in their strategic guidance to regulators and the new Regulatory Innovation Office will be empowered to drive innovation and issue targeted strategic guidance to regulators too. Pro-innovation approaches, such as regulatory sandboxes are to be promoted for high priority sectors. We have already seen a consultation on the UK’s approach to text and data mining on the context of AI.
- Building and investment in AI infrastructure: for example investment in and expansion of computational power capacity, new supercomputing facilities through the UK’s AI Research Resource, plans for allocating sovereign compute, new AI Growth Zones with streamlined planning requirements and accelerated access to clean power for AI data centres, and the creation of a new AI Energy Council to provide insight into AI energy needs and accelerate investment into energy solutions.
- Increased adoption of AI across the public sector and reduction of private sector barriers: for example determination by the Government’s Digital Centre as to when it may be appropriate to look to AI for solutions. Implementation of a Scan, Pilot, Scale approach is intended to support public sector adoption of AI. This approach will include creation of a cross-government technical horizon scanning capability. It will look to support the private sector through sharing insights, identifying opportunities for AI adoption in key industries, establishing AI sector champions to develop adoption plans with industry, providing the AI knowledge hub, entering partnerships and smoothing procurement.
- AI skills and talent: for example consideration of the AI skills gap, development of training, education, extra-curricular activities and scholarship and fellowship programmes relevant to AI.
The Government has subsequently published a series of papers addressing its approach to digital government, including its engagement with AI. This includes the State of Digital Government Review (the Review), presenting key findings on how effectively the public sector is using digital technology, including in relation to data and AI.
Also published on 21 January, the Blueprint for modern, digital government sets out the Government’s vision for a addressing some of the challenges identified in the Review (deep systemic challenges: institutionalised fragmentation; persistent legacy, cyber and resilience risk; siloed data; under-digitisation; inconsistent leadership; a skills shortfall; diffuse buying power; and outdated funding models), to achieve a digital government with a view to five outcomes (easier lives, faster growth, firmer foundations, smarter organisations, and higher productivity). The Government’s six point plan to achieve this looks to joined up public sectors, AI, digital and data infrastructure, leadership and talent, funding models and increased transparency.
The Government has also restructured various digital and AI related functions to form a digital centre within DSIT, intended to serve government departments and wider public sector. This includes the expansion of the Government Digital Service to incorporate the Centre for Digital and Data, the Incubator for AI, the Geospatial Commission and parts of the Responsible Tech Adoption Unit.
In relation AI technology specifically, the Government announced plans to use a package of AI tools to support government service delivery, called Humphrey.
We can also expect more information on the UK’s approach to AI through the Industrial Strategy’s Digital and Technologies Sector Plan, a Digital and AI roadmap in the summer and a Spring 2025 compute strategy.
The State of Digital Government Review is available here, the Blueprint for modern, digital government is available here and the AI tools press release is available here.
Matt Clifford will now act as the Prime Minister’s AI Opportunities Adviser to guide implementation of the Action Plan alongside the Department of Science Innovation and Technology’s AI Opportunities Unit.
The Action Plan is available here, and the Government response is available here.
