UK parliamentary committee says no country's AI law is "fit for purpose" -- calls for a dedicated AI Bill

The UK's Joint Committee on Human Rights -- the joint Commons-and-Lords watchdog that scrutinizes legislation for human rights impact -- published a report on September 14 calling for a dedicated AI Bill, arguing existing UK law can't handle what AI systems are already doing to people's rights. Committee chair and Labour MP Alex Sobel said no government anywhere has one: current regulatory approaches, he said, don't have "a current legislative and regulatory approach to AI that is fit for purpose."

What the committee wants

The report, closing out the committee's inquiry into AI and human rights, asks for a single independent statutory regulator -- potentially an expanded version of the UK's existing AI Security Institute -- with real enforcement teeth: mandatory pre-deployment testing and evaluation for systems that pose significant human rights risk, outright prohibitions on subliminal manipulation techniques and inappropriate biometric or emotion-inference processing, and due-diligence obligations that reach across the AI supply chain rather than stopping at the company that deploys a system. Its most unusual proposal, per outside legal analysis of the report, is giving the new regulator power to grant remedies directly to individuals harmed by AI -- without requiring them to front the cost of a lawsuit to get there.

Why now

The committee's inquiry drew on evidence of AI systems already producing the kind of harms a statute would need to address: non-consensual sexualized images of women and girls, and unconsented facial scanning. The report also flags a structural gap that shows up in most sector-specific AI regulation, in the UK and elsewhere -- liability tends to land on whoever deploys a system, not on the company that built it, leaving the actual source of a flawed model largely outside the reach of the rules meant to govern it.

The UK currently has no AI-specific statute; its approach so far has relied on existing regulators (data protection, financial services, product safety) applying their own rules to AI within their remit. Whether the government takes up a dedicated AI Bill is a political decision the report itself can't make -- committee recommendations carry weight but not force, and it lands at a moment when the US federal government is moving in the opposite direction, arguing state-level AI rules should be preempted rather than expanded.