Britain has entered a "new era of national security," the head of the country's domestic intelligence service MI5 warned on Tuesday, telling the public that threats from hostile states and terrorists are now growing — and being supercharged by artificial intelligence. In his annual threat briefing on 13 October 2026, MI5 Director-General Sir Ken McCallum said the security landscape had fundamentally changed from the relative stability of the post–Cold War years, with Russia, Iran and emerging AI capabilities now topping the agency's worry list.
The warning is not a routine annual update. McCallum said MI5's caseload for state threats — hostile activity by foreign governments on British soil — has quadrupled over the last decade, and that in 2026 the service is devoting more resources to countering state threats than at any point during the Cold War. He also revealed that MI5's terrorism and state-threat operational missions are now "pretty much of equal scale," a shift he described as enormous across recent years. Per news coverage, the speech is trending as one of the day's most significant security announcements, covered by Reuters, the BBC and CNN.
A more contested era replaces post–Cold War stability
McCallum framed the threat picture as structural rather than temporary. "A more contested era" has replaced the stability of the decades after the Cold War ended, he said, and the new environment is "not a temporary blip." The director-general called for greater national resilience and vigilance, arguing that Britain's adversaries are deliberately operating in a grey zone — through arson, sabotage, cyber-attacks, coercion and violence — that stays just below the threshold that would trigger a conventional response.
This grey-zone activity is what security professionals call hybrid warfare: coordinated hostile actions by states that blend espionage, cyber operations, sabotage, disinformation and the use of proxies, all calibrated to cause real damage while avoiding open war. McCallum's assessment is that hybrid tactics are no longer an occasional tool but the default playbook of hostile states targeting Britain — a shift with direct implications for everything from energy security to online safety, and one that mirrors Russia's continued strikes on Ukrainian energy infrastructure even after an announced energy ceasefire, as covered in reports from Ukraine.
Russia: "aggressor number one"
McCallum named Russia as the single most aggressive state actor facing Britain. He called Moscow "aggressor number one," saying it is escalating hybrid warfare while carefully staying below the threshold that would trigger a direct NATO response.
The director-general explained that Russia's ability to run traditional intelligence operations in Britain has been sharply reduced: roughly 1,000 Russian officials have been expelled from embassies across Europe since the invasion of Ukraine, degrading Moscow's professional spy networks. But rather than stopping, the Kremlin has adapted — relying increasingly on what McCallum called "native volunteers" and criminal proxies recruited to carry out operations on its behalf.
This shift matters because it makes hostile activity harder to detect and harder to deter. A professional intelligence officer can be watched, expelled or prosecuted; a criminal-for-hire recruited through encrypted messaging apps may have no visible connection to Moscow at all. The pattern matches wider European reporting of Russian-linked arson, sabotage and cyber operations across the continent, and analysts say it represents one of the most significant adaptations in state-sponsored hostile activity since the Cold War.
Iran: expanding hybrid campaign and 50 arrests
Iran was the second state actor singled out. McCallum said Tehran is also using criminal proxies — targeting political opponents, Iranian media organisations and Israeli interests in Britain — and has expanded its hybrid campaign in the last year. Operations connected to Iran have yielded more than 50 arrests in the UK, he said, with many of those operations aimed at Jewish communities.
A notable new development, according to the director-general, is Iran's ambition to move against "harder targets, like military sites" — a step up from the intimidation and surveillance operations seen previously. Iran rejects accusations of malign activity in Britain, but McCallum's figures indicate that countering Iranian operations now consumes a significant share of MI5's state-threat resources.
The warning comes amid heightened regional tensions, including Houthi attacks on Saudi Arabia and Israeli operations against Iranian interests — context that has raised concerns about whether regional conflict could spill over into hostile activity on Western soil, including attacks on Jewish and Iranian-dissident communities in Europe, as seen in the terrorism investigations already straining European security services.
AI: the threat multiplier
The most striking section of the speech concerned artificial intelligence. McCallum said AI is "supercharging the cybersecurity contest between attackers and defenders," producing tailored propaganda and personalised material for individual targets. MI5 has caught potential attackers using AI to improve reconnaissance and refine their attack methodologies, while AI-enabled drones add an entirely new challenge to physical security.
His greatest concern was explicit: terrorists using AI to develop bioweapons or even create new pathogens. While this has not yet been realised, McCallum said the risk deserves serious attention — and he announced the creation of a new AI Threat Assessment Center to track how rapidly advancing AI models could be misused by hostile actors. The announcement signals that MI5 now treats AI misuse as a standing, institutional threat category rather than a theoretical future risk.
The director-general also warned about what happens if governments fail to place sufficient "guardrails" around powerful AI: the technology could strengthen the capabilities of "comparatively low sophistication actors," effectively giving amateur attackers the kind of tools that once required state-level resources. The warning echoes recent private-sector alarms, including Anthropic's warning about self-replicating "AI worms" that could spread through smart homes — and it lands just as the most advanced AI models, like Meta's frontier Muse Wonder, are pushing further into autonomous capability.
In one of the speech's most quoted lines, McCallum said of AI itself: "Put at its simplest, human history has given us no experience of facing an actor which is smarter than we are, is capable of out-thinking us. This is genuinely novel and deserves the serious attention it is getting."
What the "equal scale" of threats actually means
The headline statistic — terrorism and state threats now roughly equal in MI5's workload — is a profound change from the post-9/11 era, when counter-terrorism overwhelmingly dominated the agency's attention. McCallum noted the terrorism threat itself has not gone away; rather, state threats have grown so fast that they now match it. The two categories increasingly overlap: the same AI tools that could help terrorists design bioweapons are also being used by state intelligence services for reconnaissance and tailored disinformation.
The director-general was careful to say the heightened threat environment is not a "temporary blip." That framing is significant: it tells policymakers, businesses and the public to plan for a permanent operating environment in which hybrid attacks, proxy operations and AI-enabled threats are routine background conditions rather than exceptional events.
Why this speech matters now
The timing of McCallum's warning is deliberate. It comes as European governments debate how to respond to escalating hybrid activity — from sabotage and arson linked to Russian proxies to drone incursions and cyber-attacks — and as AI capabilities advance faster than the guardrails around them. The creation of an AI Threat Assessment Center gives the UK a formal institutional response; the question now is whether other governments, and the tech companies building the models, move at similar speed. The sabotage threat is already before a German court, where the trial over the Nord Stream pipeline blasts has opened in Hamburg.
For the public, the speech is a rare window into how Britain's security chief sees the world: not as a collection of isolated threats but as an interconnected landscape where a hostile state, a criminal proxy, a terrorist cell and an AI model can combine into a single operation. That is the "new era" McCallum described — and, by his account, Britain is already living in it.
Those threats are not abstract: this month a police officer was shot in Nottingham city centre, a reminder of the risks faced by officers on Britain's streets.
Germany is confronting the same threat directly: a suspected Russian spy has been arrested there as Chancellor Merz warns of 'intensive' espionage.
Sources: Reuters; BBC News; CNN (reporting on MI5 Director-General Ken McCallum's annual threat speech of 13 October 2026 and subsequent media Q&A).
The response is taking shape in funding: the UK has unveiled a £600 million AI security package aimed at biosecurity and cyber threats.