Intelligence officials from the 'Five Eyes' alliance—comprising the US, UK, Canada, Australia, and New Zealand—warned on June 22, 2026, that advanced artificial intelligence models are significantly increasing the threat of cyberattacks. The alliance released a three-page statement highlighting that 'Frontier AI models' are expected to exceed industry expectations in transforming both offensive and defensive cyber capabilities. Officials emphasized that this shift is occurring over a timeline of months, rather than years.

The alert identifies specific models, including Anthropic's 'Mythos' and OpenAI's 'GPT-5.5-Cyber,' as tools that may enable users to execute complex and damaging hacks with increased speed. In response to these perceived risks, the US government recently ordered Anthropic to suspend foreign nationals' access to the Mythos model due to national security concerns. The US Cybersecurity and Infrastructure Security Agency (CISA), a cosignatory to the report, has since reduced the required deadlines for government officials to patch serious network vulnerabilities to three days.

The 'Five Eyes' statement provides broad cybersecurity guidance, recommending that organizations prioritize the swift patching of software and limit the online exposure of sensitive systems. Furthermore, the alliance encourages security teams to utilize AI to bolster defensive measures, such as the faster identification of system weaknesses and more rapid incident response times. This warning follows a period of heightened official concern regarding the dual-use nature of generative AI in digital warfare and security infrastructure protection.

Intelligence officials from the 'Five Eyes' alliance—comprising the US, UK, Canada, Australia, and New Zealand—warned on June 22, 2026, that advanced artificial intelligence models are significantly increasing the threat of cyberattacks. The alliance released a three-page statement highlighting that 'Frontier AI models' are expected to exceed industry expectations in transforming both offensive and defensive cyber capabilities. Officials emphasized that this shift is occurring over a timeline of months, rather than years.

The alert identifies specific models, including Anthropic's 'Mythos' and OpenAI's 'GPT-5.5-Cyber,' as tools that may enable users to execute complex and damaging hacks with increased speed. In response to these perceived risks, the US government recently ordered Anthropic to suspend foreign nationals' access to the Mythos model due to national security concerns. The US Cybersecurity and Infrastructure Security Agency (CISA), a cosignatory to the report, has since reduced the required deadlines for government officials to patch serious network vulnerabilities to three days.

The 'Five Eyes' statement provides broad cybersecurity guidance, recommending that organizations prioritize the swift patching of software and limit the online exposure of sensitive systems. Furthermore, the alliance encourages security teams to utilize AI to bolster defensive measures, such as the faster identification of system weaknesses and more rapid incident response times. This warning follows a period of heightened official concern regarding the dual-use nature of generative AI in digital warfare and security infrastructure protection.