Five Eyes Warn AI-Driven Cyber Threats Could Become Reality Within Months

Five-country intelligence alliance warns of a technological paradigm shift

'Cyber risk is a core business risk and a leadership responsibility'

Reuters says warning signals intelligence agencies' concern over Anthropic and OpenAI models

The Five Eyes intelligence alliance of the US, UK, Canada, Australia and New Zealand warned that artificial intelligence-driven cyber threats could become reality within months and called for immediate action.

In a joint statement released on June 22, the group said cutting-edge frontier AI models are poised to fundamentally change cyber capabilities in both attack and defense beyond current industry expectations. It said the paradigm shift could come in months, not years.

The alliance urged national leaders to act now. It called for assessments of AI-related risks, preparedness and accountability; prioritizing basic cybersecurity practices and control systems; giving cybersecurity leaders full authority and resources; and maintaining active engagement based on evolving threat indicators and guidance.

The agencies added that both organizations and society as a whole need to respond. Cyber risk is no longer just a technical issue, they said, but a core business risk and a leadership responsibility.

The statement did not spell out specific cases or evidence of AI-based cyberattacks.

Markets have viewed the warning as an extension of broad security concerns that have persisted since the so-called Mythos shock, which fueled fears that AI could fundamentally disrupt the existing security ecosystem.

Still, the warning carries added weight because intelligence agencies directly addressed cybersecurity threats stemming from AI.

Reuters said the warning showed concerns among global intelligence authorities over advanced AI models such as Anthropic's Mythos and OpenAI's GPT 5.5-Cyber had reached critical levels.

No specific model names were mentioned in the statement. Even so, it is widely seen as targeting Anthropic's Mythos, which was recently hit by export controls.

On June 12, Anthropic said the US government had issued export-control guidance fully suspending foreign access to Claude Mythos 5 and Claude Fable 5.

South Korea's government and some companies recently joined Project Glasswing, which provides access to Mythos. But the latest move by the US administration has effectively curtailed use of the model.

Park Sang-kyung, Hankyung.com reporter highseoul@hankyung.com