UK MPs call for an AI bill with powers to halt training
The report lists ten gaps in UK oversight, including no mandatory incident reporting; it is not binding, and ministers must respond by 18 January 2027 without having committed to legislate.
- Government & policy
- Safety & alignment
- Notable
The House of Commons Business, Innovation, Science and Trade Committee published its report on AI safety and security on 18 November, recorded by Parliament as HC 689, “Guardrails for growth: AI safety and security”. The details below are from The Register, which saw the report before release.
According to The Register, the committee recommends an AI Safety and Security Bill giving a watchdog power to halt high-risk training and testing, block deployments or suspend AI services if risks cannot be contained. The AI Security Institute would keep its testing role, with a separate regulator handling enforcement. The report was said to identify ten gaps in current arrangements, among them no mandatory incident reporting, limited access for independent testers and weak whistleblower protection. It also proposes compulsory reporting of incidents and near misses, and asks for legislation in the next parliamentary session and for the UK to use its 2027 G20 presidency to press for common testing and reporting norms.
The Register reported that representatives of OpenAI, Google DeepMind, Anthropic and Meta told the committee they would not release models they considered unsafe, but that the firms do not share a safety yardstick and UK law does not generally require an independent opinion before release. Committee chair Liam Byrne was quoted as saying “We cannot give today’s AI leaders unfettered freedom to behave like modern-day Dr Frankensteins”. The report also cites an OECD-based estimate that AI could make the UK economy up to £400 billion larger over a decade, which the committee used to argue that safeguards would help adoption.
Select-committee recommendations are not binding. Parliament’s record gives the government until 18 January 2027 to respond, and The Register noted it had not committed to the proposed legislation. The report followed the government’s consultation plans on agent loss-of-control, which proposed narrower rules for operators of essential services.
In the commentary
What people were saying around this time — external links, from the record's commentary rail.