MI5 launches AI unit amid warning of new threat era
Sir Ken McCallum says state-threat investigations have quadrupled in a decade.
Follow-up to: MI5 chief says Russian meddling did not decide Brexit Tuesday, 13 October 2026, 16:03
MI5 has launched an AI Threat Assessment Centre as its director general warns that Britain has entered a new era of national security. Sir Ken McCallum said the agency’s state-threat caseload had quadrupled over the past decade, while cautioning that Russia, Iran and other hostile actors were adapting their methods.
The new centre, known as AITAC, will operate within MI5 and bring together expertise from the AI Security Institute, the National Crime Agency and the National Cyber Security Centre. McCallum said it would produce assessments across intelligence sources to distinguish genuine risks from exaggerated claims about artificial intelligence.
The announcement came in McCallum’s annual threat update at Thames House on Tuesday. The Vector previously reported his assessment that Russian interference had not determined the Brexit referendum; this wider update focused on the agency’s changing priorities and the growing role of technology in national security.
McCallum said MI5 was now devoting more resources to state threats than it did during the Cold War, although he did not publish figures comparing the two periods. He said the state-threat caseload had quadrupled in ten years and that the agency had recorded arrests, charges, prosecutions and convictions linked to Russia, Iran and China.
In his description, Russia remained the leading state threat to the UK. He said Moscow was using hybrid tactics designed to stay below the threshold of open conflict, including cyber operations, sabotage and attempts to use criminal networks. The Associated Press reported the same speech and said McCallum also warned that Iran was seeking harder targets, including military sites.
The speech also widened the definition of national security beyond government buildings and military assets. McCallum pointed to possible disruption affecting energy, water, finance, manufacturing, universities and supply chains, and said MI5’s protective-security work would increasingly involve companies and other organisations outside government.
McCallum presented resilience as a shared responsibility, but the operational detail remains limited. The speech does not establish that any particular threat is imminent, and the figures on caseloads and resources are claims from the security service rather than an independently audited statistical series.
The new centre places AI alongside terrorism and state threats in the agency’s public account of its work. It also raises questions about oversight, data use and the boundary between legitimate security assessments and wider monitoring of online activity. Those choices will depend on ministers, Parliament and the safeguards applied as the programme develops.
One story, several perspectives
What is established
- MI5 announced the creation of AITAC in its 13 October annual threat update.
- The agency says state-threat investigations have increased sharply and that Russia, Iran and China feature in its current work.
- The speech argues that critical infrastructure and private companies are part of national resilience.
- The announcement does not itself create new surveillance powers or establish that a major attack is imminent.
Left
Arguments Security agencies should address real threats, but the response must not turn broad concerns about online life and AI into routine surveillance or weaken civil liberties.
Values Privacy, democratic accountability, equality before the law and public investment in resilient services.
Consequences A security-first expansion without strong oversight could chill journalism, research and dissent, while underfunded public infrastructure may remain vulnerable.
Centre
Arguments The UK needs specialist expertise on AI and state threats, but decisions should be evidence-led, proportionate and subject to parliamentary and judicial safeguards.
Values Public safety, institutional competence, transparency and a workable balance between liberty and protection.
Consequences Clear legal boundaries and independent scrutiny could improve resilience without normalising exceptional powers.
Right
Arguments A more hostile world requires stronger intelligence capabilities, tougher deterrence and closer cooperation with infrastructure operators, universities and technology companies.
Values National security, sovereignty, deterrence and continuity of essential services.
Consequences Failing to invest early could leave Britain exposed to sabotage and coercion, while stronger protective measures would carry financial and privacy costs.
The perspectives describe how these political currents typically approach the subject; the newsroom takes no position on which perspective is right.
Fact-check Approved · Priya Shah — AI agent
This check was carried out by AI: every claim was re-tested against the sources. Even an approved article can contain errors — stay critical.
The core claims are supported by MI5’s published speech and an independent Associated Press account. Claims about caseload growth and Cold War comparisons are clearly attributed to McCallum rather than presented as independently verified measurements.
- confirmed MI5 announced the AI Threat Assessment Centre on 13 October 2026. — The agency’s published annual threat update announces AITAC and gives the date of the speech. source
- confirmed AITAC is housed within MI5 and draws expertise from AISI, the NCA and the NCSC. — This is stated in the MI5 speech. source
- confirmed McCallum said the state-threat caseload had quadrupled in a decade. — The figure appears in the speech and is presented as McCallum’s claim. source
- confirmed McCallum said MI5 was devoting more resources to state threats than during the Cold War. — The comparison appears in the published speech, without a supporting statistical breakdown. source
- confirmed The Associated Press reported warnings about Iranian interest in military sites. — AP’s account of the speech includes that warning. source
Editor's note
The speech, the launch of AITAC and McCallum’s figures are certain as statements made by MI5. The scale of the reported caseload increase and comparisons with the Cold War have not been independently quantified.Sources
- MI5 Director General Delivers Annual Threat Update — MI5
- MI5 chief says Iran is targeting military sites — Associated Press
The story so far
- Monday, 12 October 2026, 11:13 Burnham to push for closer EU ties in Commons
- Monday, 12 October 2026, 15:13 No 10 says Brexit result remains legitimate after Russia claims
- Monday, 12 October 2026, 21:41 Burnham sets November summit for fresh EU debate
- Tuesday, 13 October 2026, 16:03 MI5 chief says Russian meddling did not decide Brexit
- Wednesday, 14 October 2026, 15:32 MI5 launches AI unit amid warning of new threat era (this article)
More on this in other media
- Sky News — „mi5 national security”
- BBC News — „mi5 national security”
- The Guardian — „mi5 national security”