MI5 chief warns of threats to UK from Russia, Iran and AI

The three threats operate differently but can reinforce one another. Russia is using hybrid warfare to pressure and destabilise European countries without directly fighting NATO. Iran is expanding covert activity in the UK, targeting Jewish communities and harder military sites. AI increases the speed, reach and precision of these threats.

Russia has relied on cyber-attacks, sabotage, misinformation and paid criminals after nearly 1,000 Russian spies were expelled. MI5 reported 50 arrests linked to Iranian activity in the past year. Terrorists have also tried using AI for reconnaissance and attack planning, including with AI-enabled drones.

The danger is not limited to immediate attacks. AI can tailor propaganda to individuals and help extremists improve existing methods. Sir Ken McCallum said the greatest potential risk was extremist groups using AI to harness new or existing bioweapons. MI5, GCHQ, MI6 and the AI Security Institute are working together against these threats.

Hybrid warfare is a broad term for non-military methods used to pressure or destabilise an adversary. It matters because a country can be attacked without facing conventional troops or a formally declared war. The methods may be difficult to attribute, especially when governments use criminals or other proxies.

The article lists cyber-attacks, sabotage and misinformation as examples. A cyber-attack might disrupt systems. Sabotage could damage important infrastructure. Misinformation can confuse people, weaken trust and intensify political divisions. Russia is accused of using such tactics while denying involvement despite evidence cited by the article.

This approach allows an adversary to create costs while trying to avoid direct conflict. MI5 says Russia’s traditional covert network in Britain has been weakened by expulsions, so the Kremlin is paying criminals to act for it. The threat therefore mixes technology, influence and covert physical operations.

MI5’s chief said foreign-adversary threats to the UK had quadrupled over the past decade. That increase helps explain why MI5 is now devoting more resources to state threats than it did during the Cold War. State threats include espionage, influence operations, cyber activity and covert attacks.

Iran’s activity has produced a clear recent figure. Sir Ken McCallum said there had been 50 arrests in the last year over Iran-linked activities. He said much of the aggression targeted Jewish communities, but some activity focused on harder targets, including military sites. Seven people were arrested and bailed after suspicious vans were seen near RAF Fairford.

The figures show both a broad trend and a specific operational response. State threats now consume roughly the same counterterrorism effort as terrorism. MI5 also uses the 2023 National Security Act against people working for foreign powers, including those linked to Iran, Russia and China.

Russia has stopped short of conventional conflict with NATO, according to Sir Ken McCallum. Instead, it is pursuing hybrid warfare to unsettle European countries and pressure them indirectly. This approach can create disruption while reducing the risk of a direct military confrontation with NATO members.

MI5 says almost 1,000 Russian spies were expelled after the invasion of Ukraine. With its traditional covert network weakened, the Kremlin has increasingly paid criminals to carry out attempted attacks. Cyber-attacks, sabotage and misinformation can also be conducted through proxies, making them harder to attribute and easier for Russia to deny.

The article presents this activity as a response to Russia’s weaker position, not evidence of strategic success. Sir Ken said Russia’s progress in Ukraine had been minimal and extremely costly. European leaders remain concerned that Moscow may test their commitment to supporting Ukraine through further covert or hybrid attacks.

AI can make extremist activity faster, more targeted and more adaptable. Propaganda can be tailored to individuals, which may make recruitment and radicalisation more persuasive. AI can also support reconnaissance and attack planning, lowering the effort needed to prepare harmful operations.

MI5 has caught terrorists trying to exploit AI to improve reconnaissance and attack methods. Sir Ken also mentioned AI-enabled drones as a new danger. He identified bioweapons as the most serious potential misuse, warning that extremists could use AI to harness new or existing biological weapons. That risk has not yet been realised, according to him.

The future impact is uncertain but potentially severe. AI could give violent groups capabilities once limited by expertise, time or resources. MI5, GCHQ, MI6 and the AI Security Institute are working together on these risks. Sir Ken said human history offers no experience of facing an actor smarter than humans.

The source gives a limited answer about institutional roles. It names MI5, GCHQ and MI6 together as the UK Intelligence Community, but does not set out how their responsibilities are divided. It says they are working with the AI Security Institute to address risks from artificial intelligence.

Using established public knowledge, MI5 primarily handles domestic security and counterterrorism. MI6 focuses mainly on intelligence gathered overseas. GCHQ concentrates on signals intelligence and cybersecurity. These roles overlap in practice, especially when a foreign operation targets the UK or when terrorism involves digital tools.

The article presents their cooperation as important because threats now cross traditional boundaries. AI can affect terrorism, espionage, cyber operations and foreign interference at once. It also says countering terrorism and state threats is now almost equal in scale, showing why intelligence agencies need to share information rather than work in isolation.

Distinguishing genuine threats is difficult because hostile actors deliberately hide their involvement and manipulate information. A real operation may use proxies, fake accounts or misleading claims. At the same time, political commentary and online speculation can exaggerate an incident before investigators establish what happened.

The article gives Russia’s alleged Brexit interference as an example. Sir Ken acknowledged Russian media, internet trolls and bot farms tried to influence the debate. However, he distinguished between interference being attempted and an outcome being achieved. He also warned against making such trolls look “10ft tall”, because exaggeration could serve adversaries.

This uncertainty matters for public trust and national security. Treating every suspicious claim as proof can waste resources and amplify hostile propaganda. Dismissing genuine warning signs can leave people exposed. Intelligence services therefore need evidence, careful attribution and continued investigation, especially when states use deniable proxies and online influence operations.