Dr Jaspal Kaur Sadhu-Singh discusses this week’s government announcement on the Prime Minister’s blueprint to turbocharge AI.

On the backfoot of economic challenges facing this government, the 50-point AI Opportunities Action Plan cast the die on the government’s aspirations for its place in the global AI race. In his opinion piece in the Financial Times, the Prime Minister referred to Artificial Intelligence (AI) as a “defining opportunity of our generation”.

As the US and China continue to lead AI innovation with relentless single-mindedness, the Labour Government hopes the Action Plan serves as a harbinger in this race. The wheels were set in motion after the win in the general elections in July 2024. In the Labour Party’s manifesto leading up to the elections, the tone was one to create “binding regulation on the handful of companies developing the most powerful AI models” and to set up the Regulatory Innovation Office – endorsed in the King’s Speech as “appropriate legislation to place requirements” on developers.  The cogs in the machinery have been turning leading up to the Action Plan, when in November 2024,the Department of Science and Technology published a report titled “Assuring a Responsible Future for AI” reaffirming the importance of adopting a degree of stewardship and responsibility when deploying AI systems in the delivery of public services through the use of AI assurance tools and techniques – not unlike the approach taken by the previous government’s pro-innovation approach to AI regulation policy.

Positioning the UK as the place to innovate AI, the Prime Minister has demonstrated his ability to respond to the evidence of projected exponential growth facilitated by AI. This does not allay questions on whether the economic euphoria around AI deployment, as the panacea for problems facing the UK, justifies a light-touch regulatory approach echoed by the Prime Minister’s viewthat the UK does not “need to walk down a US or an EU path on AI regulation”. Admittedly, he does not rule out regulation but instead adopting what he calls “taking a distinctively British approach that will test AI long before we regulate”. On the same note, the Action Plan alludes to introducing a regulatory regime that addresses risks.

The Action Plan comprises a list of recommendations, where Recommendations 23 through 30, under the heading of “Enabling safe and trusted AI development and adoption through regulation, safety and assurance”, set out the plans for the AI Safety Institute particularly the Institute’s role in developing AI Assurance, mentions briefly the Regulatory Innovation Office, proposes regulatory sandboxes and a degree of transparency on the part of regulators to publish the use of AI innovation tools. For the rest of the world, the UK is presenting a new case study for AI governance for countries considering the right balanced approach to regulating AI in treading the lines between innovation and responsibility – an alternative perhaps to the EU AI Act’s belt and braces approach to regulating AI.

The graduated delivery timelines in setting the Action Plans in motion will undoubtedly face challenges. Proposals appear exacting on the government’s already suffering financial resources, including the mission in Recommendation 2 to have a 20-fold increase in the UK’s supercomputing facility. Tag lines such as “Change lives by embracing AI” and “Secure our future with homegrown AI” appear impervious to the challenge of convincing the public of AI’s opportunities over its risks when AI tools are rolled out, particularly in delivering public services. The plan should not serve as a grist to the mill to the view such as that held by the Prime Minister and even perhaps technology companies, who are convinced that we are ready to harness AI, but the public may not be – rendering the best-laid plans untenable.

Dr Jaspal Kaur Sadhu-Singh is a Senior Lecturer in Law and an AI Governance expert.