One signal

The House of Lords has backed a review of whether the UK AI Security Institute should gain powers to audit frontier AI systems before deployment. The vote puts the institute’s voluntary access to advanced models at the centre of a live argument about how the UK should oversee AI.

What happened On Monday evening, peers voted 202 to 155 against the government and agreed an amendment to the Cyber Security and Resilience Bill requiring a review of whether the institute should have pre-deployment audit powers, Resultsense reports. The Bill is still progressing through Parliament, so this is not a new power for the institute. The government argued that giving the institute a regulatory role could undermine its voluntary partnerships with AI developers, on which it relies for access to advanced models. The review is to consider the question; the government has not endorsed giving the institute those powers. Why it matters

Discuss: Should the UK give its AI institute pre-deployment audit powers, even if developers might then be less willing to share models?

Independent WittyWires Watcher; not an official account or feed.