Five Eyes Warn AI Could Enable Cyberattacks in Months

Five Eyes Intelligence Agencies Warn Advanced AI Could Enable Government-Crippling Cyberattacks Within Months

Cyber security agencies from the Five Eyes intelligence alliance have issued a rare joint warning that frontier artificial intelligence models are developing at a pace that could soon grant malicious actors the capability to cripple governments, businesses, and critical infrastructure — with the threat measured in months, not years.

A Coordinated Alarm

In a joint statement published on Monday, senior cyber security officials from Australia, the United States, the United Kingdom, Canada, and New Zealand warned that advanced AI models are advancing faster than the industry anticipated. The agencies stated that these systems are “anticipated to exceed current industry expectations, fundamentally transforming both offensive and defensive cyber capabilities.”

“The timeline is not years, it is months,” the statement read. The agencies added that “cyber risk can no longer be treated as a purely technical issue. This is a core business risk and leadership responsibility.”

A Double-Edged Capability

The agencies acknowledged that AI will improve cyber defence over time. However, they warned it is simultaneously lowering the barrier for malicious actors, increasing the speed and complexity of attacks, and shrinking the window between the discovery of a vulnerability and its exploitation.

The statement urged organisations to take immediate protective steps, including:

Anthropic Models at the Centre of Debate

While the Five Eyes statement did not name any specific model or company, recent scrutiny has focused on US-based AI developer Anthropic. Earlier this year, the company determined that one of its flagship models, Mythos, was too powerful for general release and restricted access to a small group of trusted organisations.

Anthropic subsequently introduced Fable 5, a more restricted iteration of the technology. Both models were later taken offline after the US government ordered that foreign nationals be barred from accessing them, citing national security concerns.

Institutions Struggling to Keep Pace

The Five Eyes warning reflects a growing consensus among researchers, technology leaders, and security officials that AI capabilities are outpacing the capacity of governments and institutions to respond. Experts have cautioned that systems originally designed to boost productivity and strengthen defences can equally be repurposed to automate attacks and amplify the destructive reach of small, resourced groups.

The joint statement represents one of the most direct public acknowledgements to date that the window for proactive governance of AI-enabled threats may already be closing.