Five Eyes Warn of Imminent Cyber Threats from Advanced AI Models

In a joint statement issued late Monday night Sydney time, the cyber intelligence agencies of the Five Eyes alliance—Australia, the United States, the United Kingdom, New Zealand and Canada—warned that frontier AI models are poised to surpass current expectations and could reshape offensive and defensive cyber capabilities within months, lowering barriers for malicious actors and businesses.

The Guardian

Anthropic, a U.S. AI firm, released the Mythos model earlier this year and the Fable series, including community-friendly Fable 5. In June, the United States barred foreign nationals from using both models after national-security advice, prompting heightened scrutiny from allied intelligence services. The restrictions reflect growing concerns about AI misuse in cyber espionage.

The Guardian

The Five Eyes cyber agencies, Anthropic, the U.S. Department of Commerce, and the Australian government—signatory to a non-binding MOU with Anthropic—are the primary actors.

The Guardian

The alliance comprises five nations. The Five Eyes statement stresses a “months” timeline, not years. Australia’s MOU is the first under its AI plan, requiring firms to share safety-related progress and to promote safety. The alliance’s joint statement underscores the urgency.

The Guardian

Unrestricted AI could jeopardize government operations, critical infrastructure, and private-sector continuity. Framing cyber risk as a core business risk links AI threats to market confidence, urging leaders to treat resilience as a whole-of-society priority in the near term for national stability. Policymakers are urged to integrate cyber resilience into strategic planning.

The Guardian

The Five Eyes agencies called for immediate, coordinated action, labeling the threat a core business risk and leadership responsibility. The United States justified its foreign-national ban on Anthropic’s models as a security precaution. Australia’s Albanese government adopted a light-touch AI plan, emphasizing safety disclosures. Both nations stress that any regulatory approach must not stifle AI innovation.

The Guardian

Analysts warn the statement does not address potential models from other states, such as China, or from private actors. Olivia Shen noted that the next Mythos or Fable could appear imminently, complicating oversight. They call for broader international dialogue on AI governance.

The Guardian

• “act now” — Five Eyes joint statement

The Guardian

• “Frontier AI models are anticipated to exceed current industry expectations, fundamentally transforming both offensive and defensive cyber capabilities. The timeline is not years, it is months,” — Five Eyes joint statement

The Guardian

• “Cyber risk can no longer be treated as a purely technical issue. This is a core business risk and leadership responsibility.” — Five Eyes joint statement

The Guardian

• “I think we have to anticipate that the next Mythos or the next Fable is just around the corner,” — Olivia Shen, expert in national security and AI at the University of Sydney’s United States Studies Centre

The Guardian

The Five Eyes alliance will keep monitoring AI model development, while Australia and the United States consider further policy steps to balance innovation with security.

The Guardian