MI5 Chief Reveals Agency Has Caught Terrorists Seeking to Exploit AI for Attack Planning and Launches New AI Threat Assessment Centre

Analysis
MI5's disclosure moves AI misuse in UK counterterrorism from projection to documented practice, though the cases are described only in general terms and may be minor, low-impact experimentation. The announcements may serve chiefly to justify resourcing for the new Centre and the Protect mission. McCallum ranks AI-enabled bioweapons first among misuse cases but says the threat is not yet realised, so present danger lies in low-sophistication actors gaining capability uplift if model guardrails fail. The Centre gives MI5, the National Crime Agency and the National Cyber Security Centre a shared assessment channel, with the AI Security Institute reported as a partner. Energy and defence infrastructure operators now face explicit pressure to follow MI5 guidance on sabotage, drones and cyber. Confidence is high because four outlets report matching quotations from one public speech, but none is primary and all draw on a single event.