MI5 launches AI threat center as its chief warns of bioweapon risks and AI alignment

MI5, the UK's domestic security service, has launched an AI Threat Assessment Centre this year, Director General Sir Ken McCallum said on October 13 in his annual threat update.

The center, known as AITAC, is housed in MI5 and draws on experts from bodies including the UK's AI Security Institute, the National Crime Agency, and GCHQ's National Cyber Security Centre. It's producing "all-source assessments which separate the bluster from the realities," McCallum said. He compared it with the Joint Terrorism Analysis Centre, which was set up in the years after the September 11 attacks.

Bioweapons are McCallum's biggest AI worry

McCallum said AI is "already supercharging the cybersecurity contest between attackers and defenders." MI5 has caught terrorists trying to use AI to improve their reconnaissance and attack methods, he said, "with AI-enabled drones adding a new twist."

The misuse he weighs most heavily is one he described as "not yet realised": "AI's potential to change the stakes on existing or even new bioweapons." MI5, GCHQ, and MI6 are working with the AI Security Institute on those risks.

MI5 is also looking at the risk from AI itself

McCallum went beyond misuse by people. He pointed to what researchers call the alignment problem, the risk that advanced AI acts against human interests. Put at its simplest, he said, "human history has given us no experience of facing an actor which is smarter than we are." He added that the issue "deserves the serious attention it is getting, including in MI5."

The speech comes days after Washington moved on AI risk. On October 9, the White House's AI task force told AI companies that reporting incidents involving their models is mandatory. A day later, Microsoft (MSFT) CEO Satya Nadella called for an emergency brake on AI models.

Cover photo: Sir Ken McCallum. MI5, via Wikimedia Commons. Contains public sector information licensed under the Open Government Licence v3.0.