Calls for UK to introduce AI Bill to tackle risks mount

Plans for a dedicated artificial intelligence (AI) safety law, initially drafted during the final days of Keir Starmer's time in office, have resurfaced under Prime Minister Andy Burnham amid escalating public and political alarm over the existential risks associated with AI technologies.

Prior to Starmer leaving, ministers reviewed existing statutory powers and explored mandating pre-launch safety evaluations for frontier models, including allocating £115 million for an AI incident response centre and biosecurity programme. However, these efforts stalled amid the political turmoil. Upon succeeding Starmer, Burnham abolished the Department for Science, Innovation and Technology (DSIT) as he prioritised other domestic issues. AI minister Kanishka Narayan was moved to the Cabinet Office to oversee the sector without dedicated departmental resources, drawing criticism from industry experts over Whitehall fragmentation.

However, Narayan confirmed that further legislation and regulation remain actively under consideration. Proposals include barring commercial sales of unvetted models and imposing strict liability for cyber incidents. While lawmakers note that a standalone UK law is unlikely to have much effect, international coordination alongside Europe, India, Canada, and Japan would provide a more robust safe environment.

According to the Financial Times (£), proposals to create a more powerful AI regulator will be voted on at Labour conference. The vote, proposed by the Communication Workers Union, seeks to replace the AI Security Institute (AISI). The vote will be put to 7000 technology workers and will be non-binding. 

Calls for further action to rein in AI safety risks mount after Google confirmed that its Gemini AI model breached the security of three other companies in May. 

£ - This article requires a subscription.

Training Announcement: The BCS Foundation Certificate in AI examines the challenges and risks associated with AI projects, such as those related to privacy, transparency and potential biases in algorithms that could lead to unintended consequences. Explore the role of data, effective risk management strategies, compliance requirements, and ongoing governance of the AI lifecycle and become a certified AI Governance professional. Find out more.