UK AI minister to call for stronger safeguards and cyber defences
The Guardian previews Kanishka Narayan's call for frontier AI firms to do more on safety and for Britain to reduce reliance on self-policing.
A planned call for stronger oversight
UK AI minister Kanishka Narayan will call on companies developing frontier AI systems to increase their safety efforts, according to a Guardian preview published late on October 11 UTC. The report says he will announce a strengthening of Britain’s cyber defences and argue that the country cannot rely entirely on technology companies to police their own models. He is also expected to make the case for more domestic AI infrastructure.
The report comes amid concern about both unintended model actions and the security of unreleased model weights. Former UK AI Security Institute chief scientist Geoffrey Irving tells the Guardian that companies’ security would not prevent an adversarial country from stealing their models. That is his assessment, not a finding from a newly published intrusion test. The article also includes OpenAI’s response that it uses protections including encryption and continuous monitoring against model theft.
The speech is not an enacted rule
At the time of the source report, Narayan’s remarks were still forthcoming. The article does not establish that a new statutory safety regime has taken effect, specify a completed implementation schedule or confirm that the UK has joined the international call for common frontier AI controls discussed in the coverage. The immediate development is the planned ministerial intervention, not a finished legislative or technical response.