OpenAI and Anthropic summoned to parliament on fears UK AI rules ‘not fit for future’

OpenAI, Anthropic, Google Deepmind, and Meta have been called to an urgent parliamentary hearing on AI safety as MPs express concerns that Britain's rules surrounding the technology are "not yet fit for the future". Liam Byrne, chair of the Business, Innovation, Science and Trade Committee, has extended an invitation to senior representatives from all four AI giants to appear before MPs on October 13th.

The companies will be questioned on mandatory safety testing, the trigger for halting powerful model development, and the power of regulators to block a release. Byrne emphasizes trust as a core concern in their inquiry, focusing on both AI safety and the potential for workers and businesses to prosper as adoption accelerates. The committee is particularly critical of the voluntary nature of Britain's current regime, where developers ultimately decide whether the government can access their most advanced systems.

Byrne asks whether AI companies should be allowed to decide independently if their models are powerful enough to require independent safety testing, or if pre-release safety testing should be legally mandatory. The summons come after a series of warnings from industry insiders and disputes over government access to new models, with Anthropic recently releasing a model without testing it beforehand.

Byrne warns that voluntary self-regulation is unlikely to be sustainable and questions the government's powers if a company ignores safety concerns and releases a dangerous model. He also asks about legal rights for regulators to access companies' models, training data, testing methods, and safety evidence, as well as the possibility of preventing deployment, requiring further testing, or withdrawing a model that fails a safety assessment.