The UK government wants to unleash AI’s huge potential. While welcome, we need to carefully consider the risks

The UK government’s AI opportunities action plan, published on January 13 2025, outlines a vision for boosting growth and deliver services more efficiently. The plan’s muscular title is accompanied by 50 recommendations that aim to maintain the UK’s prowess as an “AI superpower”, benefiting from increased economic output, improved public services and boosted inward investment.

The recommendations, originally drafted by Matt Clifford, co-founder of the talent investment company Entrepreneur First and chair of Aria (a UK research funding agency) have been picked up wholesale by the new government and adopted as a very ambitious set of policy objectives.

As innovation director at the Surrey Institute of People-Centred AI, I argue that the fact Keir Starmer is talking about AI in such glowing terms is to be welcomed. AI represents, probably, the most powerful technology humans will ever develop.

Instead of writing detailed instructions on how a machine solves a particular problem, you show it many examples and it figures out the answer. This shift allows us to make machines that can create imagery, write text, synthesise voices and music, discover new medicines and invent new materials.

The race to create machines that are as intelligent, broadly speaking, as human beings (artificial general intelligence, or AGI), or even super-intelligences that are much smarter than us drives huge investment and resources.

Sam Altman, CEO of OpenAI, the company behind ChatGPT recently wrote: “We are now confident we know how to build AGI as we have traditionally understood it … We are beginning to turn our aim beyond that, to superintelligence.” OpenAI also published an economic blueprint for AI in America on the same day as the UK government released its plan.

The fact that statements like Sam Altman’s can now be taken seriously illustrates how incredibly important, and valuable, AI has become. The AI plan is designed to keep the UK at the forefront of technology that the UK essentially invented, with the help of pioneers like the mathematician Alan Turing and Tommy Flowers, who built the pioneering Colossus computer. It denotes a more interventionist approach to using AI, which is welcome.

The government’s plan largely focuses on laying the foundations to enable the use of AI. It recommends building the necessary computing infrastructure, freeing up data assets for AI training, improving the availability of skills and refining regulatory approaches.

It also advocates for public sector adoption and private sector investment, focusing on homegrown or “sovereign” AI. It does this by aligning UK resources and building geographic clusters of AI expertise.

We should welcome the ambition to make the UK government an intelligent customer for AI. The aggregated buying power of the UK government could attract investment and innovation.

But the challenges are immense, requiring radical changes in the commissioning, procurement and operation of IT and AI systems in the public sector. Many elements of UK government procurement are highly fragmented, and need to be better coordinated.

Moving fast

The challenges of pace and scale in global AI mean that the government will have to enact its plans extraordinarily quickly in order to be effective. AI is not a silver bullet for productivity in public services. It needs to be coupled with sensitive work on areas like culture change, workforce planning, and training.

Positive narratives need to be built for AI, countering concerns over job losses by showing people there are new roles and opportunities, with realistic pathways to attain them via education and training. Perhaps the government should go further by suggesting transition support for those industries, companies and jobs most affected by the recent developments in AI.

A welcome boost in the provision of AI courses will help, creating the required talent. However, past ambitions to improve teaching in maths and computer science have been limited by teaching capacity and specialist skills. Perhaps teaching AI with AI will become one hallmark of a leading AI nation?

The government plan aims to ensure that the UK remains in charge of its own destiny in AI, rather than simply adopting overseas technologies over which we have no control. As AI continues to become more sophisticated, it will gain in importance as a strategic national asset.

We must not underestimate the enormous challenge of competing with AI nations like the US and China where the industry driven investment dwarfs the UK government and private sector spend on AI. While competing on the global stage, the UK also needs to collaborate to create frameworks for international AI governance including the thorny issues of international data sharing and cross-border AI deployment.

Key concerns

AI safety, security and privacy remain key concerns. There will be disquiet about the plan’s proposals to share medical data with AI companies, with assurances needed to preserve privacy and to ensure that the value of this data is returned to us as citizens in the UK.

Although the plan promises to strengthen the UK AI Safety Institute, its resources are dwarfed by the scale and pace of the leading global AI companies, such as Google, Meta and OpenAI.

It’s what comes next that is important. The country’s approach to AI needs to adopt the driving principle that human beings and society benefit from AI. Discussions on productivity and efficiency can give the impression that AI is driven solely by commercial interests and reducing public service spending.

This need not be the case. With the addition of detailed roadmaps and metrics, with clear, measurable benchmarks for success in areas like talent development, infrastructure expansion, and public sector adoption, the plan becomes an impressive forward leaning piece of work.

We still need to double down on using renewable energy to power AI and find ways to make computing power more efficient. We need more ambitious training programmes and to address ethical concerns. The UK is a global leader in AI and the UK gvernment will need to deliver on every last element of the plan if it is to stay that way.

Dr Andrew Rogoyski is part of a University department that receives funding from UKRI to undertake AI research. He is an informal advisor to TechUK and a member of NatWest's Technology Advisory Board.