In London, MI5 Director General Sir Ken McCallum offered a rare and sobering public accounting of how profoundly the nature of threat has shifted beneath the surface of ordinary British life. The intelligence service now marshals more of its capacity against hostile states than it did during the Cold War, as Russia, Iran, and China pursue campaigns of sabotage, subversion, and proxy violence that blur the old boundary between war and peace. Into this already complex landscape, artificial intelligence has arrived as a force multiplier for those who wish harm, raising questions that no prior gen
MI5 Chief: Russia, Iran and AI Now Equal Terrorism Threats to UK
These are not the actions of a great power with the strategic upper hand.
So McCallum is saying state threats and terrorism are now equally resourced. That's a massive statement. What does that actually mean for how MI5 operates day to day?
It means they're splitting their attention and budget in a way they haven't since the Cold War ended. For decades, terrorism—especially after 9/11—consumed most of the oxygen. Now Russia, Iran, and the hybrid warfare they're running are demanding the same level of sustained intelligence work, personnel, and operational capacity.
But we should be careful here. He said the work is "pretty much of equal scale." That's not the same as saying they've split resources 50-50. We don't know the actual budget breakdown, and "equal scale" could mean different things depending on how you measure it.
Fair point. What about the Russia piece? He said nearly 1,000 spies have been expelled. That sounds like a win.
It is a win in the sense that MI5 has degraded Russia's traditional espionage footprint. But McCallum's argument is that Russia has adapted. They're now paying proxies and criminals to do the work instead of running their own agents. It's messier, less controllable, but it's still effective.
And we should note—he's making an inference about Russian intent and capability based on the expulsions. He's saying the Kremlin is "scrambling" and dependent on proxies. That's his interpretation of what the data shows, not necessarily the only interpretation.
What struck me was the AI piece. He said they've caught terrorists using AI for drones and reconnaissance. But he also said the bioweapon threat hasn't been realized yet. How worried should we actually be?
He's drawing a distinction between current, observable threats and potential future ones. Terrorists are already experimenting with AI-enabled drones. That's happening now. The bioweapon angle is speculative—it's a capability that could exist, but it hasn't been deployed. He's flagging it as a risk to monitor.
Right, and he's honest about the uncertainty. He said "human history has given us no experience of facing an actor which is smarter than we are." That's him admitting we're in genuinely unknown territory. The AI Threat Assessment Centre they launched is essentially saying, "We don't know what we don't know, so we're setting up a team to figure it out."
The extreme right-wing terrorism angle—he mentioned a 13-year-old. That's alarming.
It is. And it's not an outlier. He said the most marked trend over the past year has been an increase in cases where anti-immigration and anti-Muslim grievances are crossing into terrorist intent. The internet is the vector. Young people are being radicalized online, and MI5 is seeing it happen.
But we should note: he didn't give numbers on how many extreme right-wing cases there are overall, just that it's the "most marked trend." We know the terrorism threat level is severe, and he's saying a significant part of that is driven by this category. But the actual scale—how many people, how many active plots—that's not in his public statement.
And the RAF Fairford incident—he said it's "not unique." So Iran is running multiple operations against UK military sites?
That's what he's indicating. Fifty arrests in the past year linked to Iranian activities. Some of it is targeting Jewish communities, but he's also pointing to "harder targets, like military sites." RAF Fairford is the one that became public, but he's suggesting there are others.
Though he didn't specify how many other incidents there have been or how many were actually successful versus disrupted. "Not unique" tells us it's a pattern, but the full scope isn't clear from what he said publicly.
Il Polso
- MI5 now spends more on countering state threats than at any point in the Cold War, a reordering that signals foreign adversaries have become as dangerous as terrorist networks.
- Russia leads the field as 'aggressor number one,' running hybrid warfare operations — cyber-attacks, sabotage, disinformation — across Europe while nearly 1,000 of its spies have been expelled from the UK, forcing Moscow to rely on criminals and paid proxies.
- Iran has mounted 50 arrests' worth of operations on British soil in a single year, targeting Jewish communities and military installations, with the RAF Fairford incident — three suspicious vans, seven arrests — standing as the most visible example.
- Extremist groups are already weaponizing AI for reconnaissance and drone integration, and McCallum warns the technology could eventually lower the barrier to bioweapon development — a risk MI5's new AI Threat Assessment Centre is racing to understand.
- Domestic terrorism has shifted in character: the threat level is at severe for the first time in five years, extreme right-wing cases are rising sharply, and MI5 recently investigated a radicalized 13-year-old whose case helped prevent a lethal attack abroad.
In London, MI5 Director General Sir Ken McCallum offered a rare and sobering public accounting of how profoundly the nature of threat has shifted beneath the surface of ordinary British life. The intelligence service now marshals more of its capacity against hostile states than it did during the Cold War, as Russia, Iran, and China pursue campaigns of sabotage, subversion, and proxy violence that blur the old boundary between war and peace. Into this already complex landscape, artificial intelligence has arrived as a force multiplier for those who wish harm, raising questions that no prior generation of security thinkers has had to answer. The warning is not of a single crisis but of a permanent reconfiguration — one in which the adversaries are many, their methods are evolving, and the familiar categories of danger no longer hold.
Sir Ken McCallum took the stage in London to deliver something unusual for an intelligence chief: a frank, public reckoning with how completely the threat environment has changed. MI5, he announced, now devotes more resources to countering hostile states than it did throughout the entire Cold War. Terrorism and state-level aggression have reached roughly equal scale inside the service's workload — a reordering driven by a decade in which foreign threats have quadrupled.
Russia occupies the center of this new picture. McCallum called it 'aggressor number one,' describing a sustained campaign of hybrid warfare — cyber-attacks, sabotage, disinformation — calibrated to destabilize European nations without triggering direct NATO confrontation. The expulsion of nearly 1,000 Russian spies since the Ukraine invasion has degraded Moscow's traditional intelligence apparatus, leaving it dependent on criminals and paid proxies. McCallum was pointed in his assessment: 'These are not the actions of a great power with the strategic upper hand.'
Iran has been no less active. Fifty arrests linked to Iranian operations were recorded in Britain over the past year, many targeting Jewish communities and military sites. The discovery of three suspicious vans near RAF Fairford in Gloucestershire — resulting in seven arrests — drew particular attention. Prime Minister Andy Burnham described the incident as almost certainly directed by Iran's Revolutionary Guard, and McCallum confirmed the airbase was not an isolated target.
Artificial intelligence has entered this landscape as a complicating force. Extremist groups are already using it to refine reconnaissance and integrate drones into attack planning — MI5 has caught terrorists attempting exactly this. The longer horizon concern is bioweapons: McCallum warned that AI could eventually enable the development or modification of biological agents, though he stressed the threat has not yet materialized. 'Human history has given us no experience of facing an actor which is smarter than we are,' he said. A new AI Threat Assessment Centre has been established to separate credible risk from speculation.
At home, the terrorism threat level was raised to severe in April for the first time in five years. Islamist networks continue to inspire and direct attacks, but the sharpest trend of the past year has been the rise of extreme right-wing terrorism, driven by anti-immigration and anti-Muslim grievances spreading through the internet and skewing toward younger perpetrators. MI5's investigation of a radicalized 13-year-old recently produced intelligence that allowed an international partner to prevent a lethal attack.
McCallum also addressed Russian interference in the 2016 Brexit referendum, acknowledging that media outlets, trolls, and bot farms had attempted to shape the debate — but cautioning against overstating their reach. 'I am a little bit anxious that we don't play into our adversaries' hands by making these trolls look 10ft tall,' he said. His broader message was unambiguous: the security environment has been fundamentally redrawn, and MI5's response — more resources, new legal tools, and a new centre for AI threats — reflects a service that knows it.
Sir Ken McCallum stood before an audience in London and delivered a stark recalibration of British security priorities. The head of MI5 announced that the intelligence service now devotes more resources to countering state threats than it did during the entire Cold War—a seismic shift in how the country allocates its defensive capacity. The work of tackling terrorism and state-level aggression, he said, had reached roughly equal scale, a reordering that reflects a decade in which threats from foreign adversaries have quadrupled.
Russia sits at the center of this new calculus. McCallum named it "aggressor number one," describing a campaign of hybrid warfare—cyber-attacks, sabotage, disinformation—designed to unsettle European nations without crossing into conventional military conflict. The Kremlin has stopped short of direct confrontation with NATO, he noted, while its full-scale invasion of Ukraine grinds forward at minimal territorial gain and horrific cost. What makes Russia's position precarious, in McCallum's assessment, is the damage already inflicted on its traditional espionage apparatus. Nearly 1,000 Russian spies have been expelled from the UK since the invasion began, leaving Moscow dependent on paying proxies and criminals to carry out operations on its behalf. "These are not the actions of a great power with the strategic upper hand," McCallum said.
Iran has mounted its own escalating campaign across British territory. In the past year, MI5 recorded 50 arrests linked to Iranian activities, much of it directed at Jewish communities and military installations. The incident at RAF Fairford in Gloucestershire—where three suspicious vans were discovered near the airbase—exemplifies this pattern. Seven people were arrested and bailed by counter-terrorism police in connection with that case. Prime Minister Andy Burnham has said the incident was "almost certainly directed by Iran's Islamic Revolutionary Guard." McCallum confirmed the airbase was not an isolated target. "We have seen Iranian interest beyond RAF Fairford in the UK," he told reporters.
The emerging threat landscape now includes artificial intelligence in ways that complicate both attack and defense. Extremist groups have already begun exploiting AI to sharpen their reconnaissance, refine attack methodologies, and integrate the technology with drones. MI5 has caught terrorists attempting precisely this. The more distant but graver concern is bioweapons. McCallum warned that AI has the potential to enable the development or modification of biological weapons, though he stressed this threat has not yet materialized. "Put at its simplest, human history has given us no experience of facing an actor which is smarter than we are," he said, acknowledging the unknowable dimensions of the risk. To address this, MI5 has established an AI Threat Assessment Centre tasked with separating credible threat from speculation.
Domestic terrorism remains a persistent pressure point, and the character of that threat has shifted. In April, the UK raised its terrorism threat level to severe for the first time in five years, signaling that an attack is highly likely. Islamist groups including Al Qaeda and Islamic State continue to direct and inspire attacks. But the most marked trend over the past year has been the rise of extreme right-wing terrorism. McCallum described a pattern in which anti-immigration and anti-Muslim grievances are tipping into terrorist intent. This category of threat skews young and spreads through the internet. MI5 recently investigated a radicalized 13-year-old in the UK; that investigation yielded intelligence that enabled an international partner to prevent a lethal attack on their soil.
When asked about Prime Minister Burnham's recent assertion that Russia interfered with the Brexit referendum, McCallum acknowledged it as established fact but drew a careful distinction. Russian media outlets, internet trolls, and bot farms had attempted to influence the 2016 referendum debate—that is public record. But he emphasized the referendum result itself was not in dispute and expressed concern that overplaying Russian interference could inadvertently amplify the reach of those operations. "I am a little bit anxious that we don't play into our adversaries' hands by making these trolls look 10ft tall," he said.
The National Security Act, passed in 2023 to target individuals working against UK interests on behalf of foreign powers, has become a heavy-use tool in this expanded security architecture. It has been deployed against threats from Russia, Iran, and China alike. McCallum's message was unambiguous: "Those who seek to cause mayhem on British streets—whether hired thugs here or their paymasters overseas—will meet an unflinching response." The rebalancing of MI5's resources, the identification of state actors as co-equal threats to terrorism, and the acceleration of AI-enabled risks all point toward a security environment fundamentally different from the one that prevailed even five years ago.
Citazioni salienti
Put at its simplest, human history has given us no experience of facing an actor which is smarter than we are.— Sir Ken McCallum, MI5 chief, on artificial intelligence threats
I am a little bit anxious that we don't play into our adversaries' hands by making these trolls look 10ft tall.— Sir Ken McCallum, on Russian interference in Brexit and the risk of amplifying disinformation