The head of Britain's domestic intelligence agency, MI5, said Tuesday that artificial intelligence is "supercharging the cybersecurity contest between attackers and defenders" and that his greatest AI concern is its potential to raise the danger of biological weapons, Reuters reported. Coming from the security chief of a close U.S. ally, the assessment adds weight to calls for security controls on advanced AI in Britain and allied countries, including the United States.

What Happened

Director General Ken McCallum gave his annual speech on the threats facing Britain. According to Reuters, he said terrorists are using AI to improve reconnaissance and attack planning, and that he and his colleagues are "particularly focused" on how AI could change the stakes of existing or new bioweapons. "These risks are real," he said, as quoted by Reuters.

McCallum also said threats from foreign states have quadrupled over the past decade and that MI5 now devotes more resources to them than it did during the Cold War.

RelatedAnthropic Asks Appeals Court to Rehear Ruling Upholding Pentagon's Supply-Chain Risk Label

He called Russia "aggressor number one," Reuters reported, citing an attempted drone attack on a cargo plane at a German airport and a cyberattack on Polish energy infrastructure. Russia denies hostile acts in Britain and elsewhere; its London embassy called Western attributions "convenient," according to Reuters.

On Iran, McCallum described a shift this year toward harder targets such as military sites, and said Iran-linked operations had led to more than 50 arrests in Britain over the past year. He declined to discuss a suspected plot against RAF Fairford, an air base used by U.S. bombers, citing an ongoing investigation. Reuters reported that Prime Minister Andy Burnham had said a day earlier that Tehran was "highly likely" behind it; Iran has rejected involvement.

The speech text was not available from MI5's website when this article was prepared, so this account relies on Reuters' reporting.

Why It Matters

McCallum's speech is the agency's main public threat assessment, and he placed AI alongside Russia and Iran as part of the same threat picture. His framing of AI as speeding up both attack and defense means the advantage goes to whichever side adopts it more effectively. For organizations, that puts pressure on how quickly defenders can use AI tools to find and fix weaknesses compared with how quickly adversaries can use them to find and exploit them.

His focus on biological weapons is significant for AI developers and the companies that deploy their models. It signals that intelligence agencies regard the misuse of capable AI systems for weapons work as a live national security issue, which tends to drive stricter testing expectations and controls on model access.

The figures on state threats matter for boards in sectors that foreign governments target. The examples McCallum cited, energy infrastructure and logistics, show that hostile states are willing to go after civilian companies as well as government targets.

What to Watch

Reuters did not report new figures on AI-enabled plots, and MI5 has not said how it measures AI's role in the threats it tracks. Watch for whether British lawmakers draw on the speech in debates over cyber resilience legislation and AI risk rules, and whether allied agencies issue similar warnings on AI and biological threats. The speech came a day after Britain's AI minister called a breach of national cyber defenses likely.