Credit: The wub / CC BY-SA 4.0 via Wikimedia Commons A cross-party committee of MPs and peers wants a dedicated UK AI law and a single regulator with enforcement powers. On the same day, OpenAI said it would back binding UK rules for the few companies building the most powerful AI. That includes OpenAI itself. Tom Duff Gordon is OpenAI’s head of policy for Europe, the Middle East and Africa.
The “U.K. should act now to set up a stronger set of safety rules,” he told Joseph Bambridge of Politico. According to Politico, it is the first time OpenAI has called for binding UK rules on frontier AI. He said Prime Minister Andy Burnham’s government should “take advantage of this political window which is clearly opening up”. What the committee wants The Joint Committee on Human Rights published its report, Human Rights and the Regulation of AI, on 14 September.
It found that UK law covers some AI systems but leaves serious gaps. No single body coordinates AI regulation, so people harmed by AI can struggle to find who is responsible. The committee wants an AI bill with stricter obligations for higher-risk systems. The bill would also ban uses the committee considers incompatible with human rights.
It also wants mandatory transparency across the AI lifecycle and an independent oversight body set up in law. The report argues that more responsibility should sit with the developers of foundation models, not only the companies that deploy them. It also wants a timeline for ratifying the Council of Europe’s AI convention. The report lists risks it says existing law does not cover.
Police facial recognition scanned around 3 million faces in the UK without consent between January and October 2025, the committee estimates. It also cites sexually explicit deepfakes and workplace systems that flagged drivers for discipline after they stopped at traffic lights. Alex Sobel is the Labour MP who chairs the committee. “Nowhere in the world, including the UK, has a current legislative and regulatory approach to AI that is fit for purpose,” he said. The report also points to the limits of the AI Security Institute (AISI).
The institute tests frontier models before release, but it has no regulatory authority and cannot demand access to them. Last week, Anthropic skipped UK pre-release tests for Mythos 5.1, according to the Financial Times. What OpenAI is asking for Duff Gordon described what OpenAI would accept. “OpenAI is in a position to support legislation in the U.K. to establish durable, mandatory, capability-based requirements for frontier AI,” he told Politico. Any law should focus narrowly on national security and cybersecurity risks, he said.
It should cover only the small number of labs and models at the frontier. He wants compulsory independent testing by third parties and rules for monitoring and reporting safety incidents. He also said AISI should play a central role. The Guardian quoted him separately. “We support stronger UK rules for the handful of companies, including OpenAI, developing the most powerful AI systems,” he said.
OpenAI said the rules should not cover startups working on less powerful systems. The company has taken a similar line in the US. Chief global affairs officer Chris Lehane wrote a blog post on 9 September. In it, he said OpenAI wants to work with Congress on “mandatory, capability-based national AI safety regulation”.
Pressure on the government Labour’s 2024 manifesto promised binding regulation for the handful of companies developing the most powerful AI models. The government has not yet introduced that legislation. It has relied instead on sector regulators and voluntary testing through AISI, Politico reported. The calls follow a busy weekend.
Anthropic chief executive Dario Amodei urged the industry to slow down and OpenAI’s Sam Altman backed the idea. In August, AISI testers caught OpenAI and Anthropic agents acting without authorisation. OpenAI has also disclosed that its own models broke out of testing to hack Hugging Face. A government spokesperson told the Guardian: “We should not assume that the existing framework will always be sufficient as AI capabilities develop.
Any future measures will be evidence-based and focused on the risks we need to address.” A warning against overreaction Business secretary Jonathan Reynolds struck a different tone on 15 September. He told BBC Radio 4’s Today programme that people should not get “hyperbolic” about AI risks, the Guardian reported. He warned that regulation which cut the UK off from frontier developments “would obviously make us less safe”. First secretary Louise Haigh was due to tell the TUC conference the same day that ministers must “heed the warnings” from the people building the technology.
Parliament is also turning to AISI itself. Liam Byrne, the Labour MP who chairs the business and trade committee, has asked the institute to give evidence at a hearing next month, the Guardian reported.
Leave a Reply