Five Eyes warns new AI models pose urgent cyber risk within months

The world’s most powerful intelligence alliance just told everyone to brace for impact. The Five Eyes, the spy partnership linking the US, UK, Canada, Australia, and New Zealand, issued a joint statement on June 22 warning that frontier AI models will dramatically escalate cyber attack capabilities “within months rather than years.”

What the Five Eyes actually said

The statement, titled “The AI shift in cyber risk: why leaders must act now,” carries the signatures of the heads of six major cyber and intelligence agencies: the Australian Cyber Security Centre, Canadian Centre for Cyber Security, New Zealand’s NCSC, the UK’s NCSC, the US NSA, and CISA.

The core message: frontier AI models are anticipated to exceed current industry expectations. The agencies are urging organizations across every sector to prioritize cyber resilience immediately, recommending secure-by-design approaches, accelerated patching of vulnerabilities, reduced attack surfaces, and, notably, adopting AI for defensive purposes.

The Anthropic connection

The statement didn’t arrive in a vacuum. Around June 12, the US Commerce Department ordered Anthropic to restrict global access to its advanced AI models, including one referred to as “Mythos.” That’s a remarkable move, effectively treating a private company’s AI product as a national security concern requiring export-style controls.

The Five Eyes agencies had also issued guidance in May 2026 cautioning against the rapid deployment of agentic AI systems. Those are AI models capable of taking independent actions rather than just generating text or images. The concern was that rushing these systems into production could amplify existing vulnerabilities and introduce new vectors for misbehavior, whether intentional or accidental.

So the June 22 statement represents the third major escalation in roughly six weeks. May guidance on agentic AI risks. June restrictions on Anthropic’s models. Then a historic joint declaration from all five nations.

Why crypto should pay attention

Crypto protocols are, by design, high-value targets with publicly visible code and enormous financial incentives for attackers. DeFi exploits have already drained billions from the ecosystem over the past few years using relatively conventional attack methods. The Five Eyes’ emphasis on accelerated patching timelines is particularly relevant here. In DeFi, the window between vulnerability disclosure and exploit is often measured in hours, not days.

On the regulatory front, the Anthropic restrictions hint at a broader willingness by governments to impose access controls on advanced AI capabilities. For crypto projects integrating AI features, whether for trading, risk management, or autonomous agents, this creates a new layer of compliance uncertainty. If governments start restricting which AI models can be accessed and by whom, projects built on those models face potential disruption.

The Five Eyes explicitly recommended using AI defensively, which is essentially an endorsement of the technology when deployed responsibly.

Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.