MI5 chief puts AI among its threats and launches a centre to assess it

MI5 Director General Ken McCallum spoke about AI in his annual threat speech on 13 October. He said MI5 has caught terrorists trying to use it to improve their reconnaissance and attack methods. It hasn't happened yet, he added, but the misuse he gives most weight is AI changing the stakes of existing or new bioweapons. He announced the AI Threat Assessment Centre (AITAC), launched in 2026 with expertise from British bodies such as the AISI and the NCSC. He also mentioned the "alignment problem": we have never faced an actor smarter than we are.

On 16 October, the BBC's AI Decoded discussed it. Researcher Roman Yampolskiy called for an international ban on superintelligence. Technology futurist Bernard Marr replied that there is no proof that a more intelligent AI would have goals different from ours.