Five Eyes Alliance Warns AI‑Powered Cyber Attacks Will Arrive Within Months

The intelligence partnership of the United States, United Kingdom, Australia, Canada and New Zealand (the Five Eyes) issued a joint advisory stating that frontier artificial‑intelligence models capable of automatically generating and deploying multi‑stage cyber exploits are expected to become widely accessible in a matter of months, not years. The directive urges governments and private enterprises to accelerate patch cycles – for example, the U.S. Cybersecurity and Infrastructure Security Agency reduced its federal vulnerability‑patch window to 72 hours – and to adopt aggressive hardening measures such as shrinking attack surfaces, eliminating legacy systems, enforcing strict multi‑factor authentication and assuming a breach posture. In the United States, Executive Order 14409 establishes an AI‑cybersecurity clearinghouse, mandates classified benchmarking of “covered frontier models,” and calls for rapid sharing of defensive AI tools with critical‑infrastructure operators like hospitals and community banks. The warning emphasizes that AI will compress the time needed for sophisticated intrusions, expanding the pool of potential adversaries and increasing risk to essential services.