UK Parliament's human rights committee calls for new AI law and dedicated regulator

A Stephenson Harwood edition of its Neural Network newsletter, covering AI developments in August and September 2026, includes a section on a report from Parliament's Joint Committee on Human Rights. The committee published a 100-day report on 14 September 2026 urging new legislation on the human rights risks of fast AI adoption and a dedicated regulator to coordinate the response.

The committee described the current legal framework as patchy and confused and named three main problems. First, the Equality Act 2010 covers only people receiving services and leaves out business-to-business settings, so bias in AI may go unaddressed. Second, gathering web data to train models probably breaches UK GDPR, and larger volumes of personal data make re-identification of anonymized information easier. Third, limited transparency means people often cannot tell whether AI was used and so cannot challenge it, leaving gaps in the right to an effective remedy.

The report makes two main proposals. One is an AI bill implementing in full the international framework convention on AI covering human rights, democracy and law, with obligations scaled to risk and pre-approval for especially high-risk models. The other is a single statutory independent oversight body with broad powers to coordinate, monitor and enforce. The AI Security Institute would also be put on a statutory footing, able to review powerful models and advise the regulator on blocking or withdrawing systems that pose serious rights risks. The newsletter notes that no AI bill progressed under the Starmer government, and that Andy Burnham made Kanishka Narayan a cabinet-level AI Minister on 20 July 2026.