MI5 Warns of AI Bioweapon Risks, Launches New Threat Center

Sir Ken McCallum, Director General of MI5, warned on October 13, 2026, about the “real risks” of artificial intelligence (AI) developing biological weapons and announced the establishment of a new AI Threat Assessment Centre[5][8]. This new center is a direct institutional response to concerns that rapidly advancing AI could enable the creation of new pathogens or give terrorists access to bioweapons, raising the potential for bioterrorism and global pandemics.

What the records show

Sir Ken McCallum delivered his annual threat update at Thames House in London, identifying AI’s potential to “change the stakes” for existing or new bioweapons as the primary misuse concern[5][6][8]. He said AI could offer a “capability uplift” to “comparatively low-sophistication actors” if technological “guardrails” fail[5][7]. McCallum also noted that AI is “already supercharging the cybersecurity contest between attackers and defenders” and can produce “propaganda and radicalising material tailored to the individual”[6][7][8]. He cited instances of “terrorists seeking to exploit AI to improve their reconnaissance and attack methodologies” with “AI-enabled drones adding a new twist”[7][8][12]. He described this period as a “new era of national security,” stating MI5 now dedicates more resources to state threats than during the Cold War[8][16].

MI5’s new AI Threat Assessment Centre (AITAC) will be an analytical hub, combining expertise from the UK AI Safety Institute (AISI), the National Crime Agency (NCA), and the National Cyber Security Centre (NCSC)[8][10][12]. Its stated goal is to produce “all-source assessments” to “separate the bluster from the realities” of AI threats, which will aid informed decision-making[8][16]. This is MI5’s third analytical center, alongside those for terrorist and state threats[16][17].

Dr. Tom Inglesby, Director of the Johns Hopkins Center for Health Security, expressed similar concerns on October 14, 2026, regarding the lack of regulation in the field[9]. He stated that while some AI companies take precautions, “the majority are not announcing what, if anything, they are doing”[9]. Inglesby observed that “no formal government policies in place in any country yet to protect against this kind of misuse”[9]. He advocates for “formal independent evaluations of new AI models before they are released” and restricting “high-risk biological data”[9]. In November 2023 testimony to Congress, Dr. Inglesby highlighted “dual-use risks of AI and biosecurity (AIxBio),” specifically mentioning AI’s potential to “accelerate or simplify the reintroduction of particularly dangerous extinct viruses” or to “enable, accelerate, or simplify the creation of entirely new biological constructs that could start a new pandemic”[4].

The timeline

On October 13, 2026, Sir Ken McCallum delivered his annual threat update and announced the AI Threat Assessment Centre[8]. The following day, October 14, 2026, Dr. Tom Inglesby publicly commented on AI misuse and regulatory gaps[9]. Dr. Inglesby had previously testified to the U.S. Congress in November 2023 on AI and biosecurity risks[4].

What changed and for whom

With the establishment of the AI Threat Assessment Centre, MI5 is creating a coordinated, multi-agency approach to analyze and mitigate AI-related national security risks[8][10]. This move represents a direct institutional response by the UK’s domestic intelligence agency to what it calls a “new era of national security”[8][16]. For citizens, this means a heightened governmental focus on potential AI-enabled threats, ranging from cyberattacks and propaganda to bioweapons[6][7][8].

What the subject has said publicly

Sir Ken McCallum stated that the primary misuse concern is “AI’s potential to change the stakes on existing or even new bioweapons”[5][8]. He added, “These risks are real”[6][8]. Dr. Tom Inglesby commented that “no formal government policies in place in any country yet to protect against this kind of misuse” of AI in biology[9]. He also noted that “the majority [of AI companies] are not announcing what, if anything, they are doing” to prevent misuse[9].

What we could not confirm

We could not confirm specific details regarding the guardrails AI companies are implementing or considering. The exact timeline for when formal government policies might be established to mitigate AI bioweapon risks in any country remains unclear.