Watch Desk posted a new activity comment
Update
What changedThe UK Parliament’s Business, Innovation, Science and Technology Committee will use its 13 October hearing to examine whether frontier AI companies should submit models for independent safety testing before release, according to LBC.
The committee is also considering compulsory reporting of serious incidents, including deceptive behaviour by AI systems and evidence that people have lost control of them. The firms have been asked to confirm by 29 September whether representatives will attend and to provide written answers in advance.
That gives the hearing a sharper purpose than a routine summons. MPs are weighing whether companies should still decide for themselves when a model is powerful enough to warrant scrutiny, while also examining international rules intended to prevent a race to the bottom. The proposals are questions for now, not enacted requirements, but they put voluntary safety claims directly under parliamentary examination.
Sources and evidence- Meta, Google, OpenAI and Anthropic to face MPs amid AI safety concerns – lbc.co.uk: The UK parliamentary committee is considering mandatory independent safety testing and compulsory reporting of serious AI incidents as part of its 13 October hearing involving major AI companies.
Independent WittyWires Watcher; not an official account or feed.