Source: Global Legal Insights

MPs and peers on the Joint Committee on Human Rights publish a report on 14 September calling for a dedicated UK AI law that would make designers, developers, suppliers and users subject to duties calibrated to their role and the risk of the system, and would give a new regulator the power to prevent unsafe high‑risk systems reaching the public. High‑risk systems would require approval before provision or use, organisations would have to disclose and explain AI’s contribution to important decisions about people, and some uses—identified as subliminal manipulation, emotional inference and inappropriate profiling or biometric use—could be banned following public consultation.

The committee argues existing laws place responsibility on the user even where risks are built into a system and cites examples such as a delivery driver unable to log on due to facial recognition failure, workers flagged after stopping at traffic lights, AI profiling of prisoners and public face‑scanning. The report recommends an independent oversight body to investigate complaints, set enforceable codes, maintain a public incident record, test and block high‑risk systems and order remedies; it also proposes mandatory submission of powerful models to the AI Security Institute for testing, noting the institute currently relies on voluntary developer access. The committee contrasts its proposals with the government’s stance that existing regulators govern AI “at the point of use” and urges new legislation to manage risk across the AI supply chain and lifecycle.