Britain’s intelligence chief says terrorists are seeking to exploit artificial intelligence, as a new assessment centre examines the technology’s implications for national security.
LONDON: Artificial intelligence is intensifying the security threats facing Britain, MI5 director general Sir Ken McCallum warned on Tuesday, 13 October, as he outlined concerns ranging from cyberattacks and extremist propaganda to the potential misuse of biological research.
Speaking at MI5’s headquarters, Thames House, McCallum said the technology was already affecting the contest between cyber attackers and defenders. His annual threat assessment also addressed hostile activity attributed to foreign states and continuing terrorist threats.
Terrorists seeking to exploit AI
McCallum said MI5 had detected terrorists seeking to use AI to improve reconnaissance and methods of attack. He also identified AI-enabled drones as an additional challenge.
The technology can generate propaganda and radicalising material tailored to individuals, he said, placing its influence on online content alongside its potential role in preparing attacks.
Biological weapons flagged as a future risk
McCallum identified the potential use of AI to assist with existing or new biological weapons as the misuse scenario that concerned him most.
He explicitly described that scenario as a risk that had not yet been realised. His warning therefore concerned what the technology might enable, rather than an announcement that MI5 had uncovered an AI-developed biological weapon.
New centre assesses national security threats
McCallum announced that the AI Threat Assessment Centre, known as AITAC, had been launched in 2026.
MI5 says the centre is based within the agency and brings together expertise from bodies including the AI Security Institute and the National Cyber Security Centre. Its role is to assess AI-related national security threats for government departments.
The centre gives those departments a dedicated source of assessments as they consider the security implications of rapidly developing AI systems.
Warning comes amid wider state threats
The AI assessment formed part of a broader warning about Britain’s security environment. McCallum said MI5’s caseload involving hostile states had quadrupled over the past decade, Sky News reported.
He accused Russia of escalating activity against Britain and its allies through tactics that stop short of conventional warfare. He also said Iran was seeking to expand its activities to harder targets, including military sites.
Russia and Iran deny accusations of hostile activity in Britain, Reuters reported.
Technology firms outline response to safety concerns
In a separate development on 13 October, representatives of OpenAI and Anthropic told MPs they would act on evidence from British intelligence agencies that their AI systems were unsafe, according to Sky News.
The broadcaster reported that the companies’ representatives said they could halt deployment or remove systems from the market where they retained control of the models.
Both also highlighted a practical limitation: that control is not always available when a model is hosted by another organisation. Their commitments were therefore conditional on the companies being able to intervene in how the affected system was deployed.
