The Five Eyes intelligence alliance warns that leading AI systems are advancing so quickly they could outpace current cybersecurity defenses within months. The alert was prompted in part by Anthropic's claims that its Mythos models can identify software vulnerabilities at an unprecedented level. The agencies urge governments and businesses to integrate AI into defensive operations, modernize legacy systems, restrict access to critical infrastructure and improve incident preparedness. Anthropic has limited access to Mythos 5 and Fable 5 after a U.S. national security directive.
Five Eyes Warns Advanced AI Could Outpace Cybersecurity Within Months — Urges Immediate Action
AI on pace to bypass cybersecurity systems soon, "Five Eyes" spy partners warn
The Five Eyes intelligence alliance — representing the security services of the United States, United Kingdom, Australia, Canada and New Zealand — warned that the most advanced artificial intelligence systems are evolving fast enough to surpass current cybersecurity defenses in a matter of months.
The alert follows claims by AI startup Anthropic that its Mythos family of models demonstrated unprecedented abilities to identify software vulnerabilities. Security agencies say this rapid progress means assumptions about cyber risk can become outdated far more quickly than organizations expect.
"The rapid pace of frontier AI development means cyber risk assumptions can become outdated in months, not years," the joint Five Eyes statement said, adding that AI "lowers barriers for malicious actors and increases the speed and complexity of attacks."
The agencies stressed that breaches are inevitable, but preparedness can limit impact and prevent escalation into broader operational or financial crises. To strengthen defenses they recommended a combination of practical measures aimed at both reducing attack surfaces and improving response capabilities.
Recommended Steps for Organizations
Integrate AI Into Security Operations: Use AI tools to detect, triage and respond to threats faster while understanding the risks these tools may introduce.
Modernize Legacy Systems: Update or replace outdated infrastructure that can be easily exploited by automated tools.
Restrict And Monitor Access: Limit privileges to critical systems, enforce strong authentication, and continuously monitor for unusual behavior.
Improve Incident Preparedness: Run tabletop exercises, maintain clear escalation paths, and plan for rapid containment to avoid widespread disruption.
Chris Krebs, former director of the U.S. Cybersecurity and Infrastructure Security Agency, described the guidance as a practical checklist organizations can adopt to become "harder targets" and more resilient in the face of what he called a looming "vulnerability tsunami."
This month Anthropic restricted access to Mythos 5 and a limited variant called Fable 5 after receiving a U.S. national security directive that barred foreign nationals from using the models. The intervention highlights rising tensions between rapid AI development and national security concerns, even as the White House has signaled interest in rolling back some AI restrictions.
In short, the Five Eyes advisory is a reminder that defenders must move quickly: adopting stronger controls, leveraging AI defensively, and preparing to respond when breaches occur.
Help us improve.
