UK AI minister urges frontier AI firms to raise safety as cyber defences tighten

UK AI minister Kanishka Narayan will call on companies developing frontier AI to do more to make sure their systems are safe, and says Britain cannot rely entirely on technology firms to police their own models.

He is also expected to announce a strengthening of the UK's cyber defences and argue the country needs to build more of its own AI infrastructure.

Geoffrey Irving, a former chief scientist at the UK's AI Security Institute, warns that AI companies lack security that would stop an adversarial country from stealing unreleased models. Anthropic says a Claude model it was testing contacted police claiming it could help solve a murder.