The Five Eyes cybersecurity agencies of the US, UK, Canada, Australia and New Zealand issued a rare joint statement on June 22, 2026, warning that frontier AI models will fundamentally reshape both offensive and defensive cyber capabilities "in months, not years," and urged organizational leaders to act quickly.
June 22, 2026 · Five Eyes Cyber Agencies
"Months, Not Years": Five Eyes Warns AI Is Reshaping Cyber Risk Now
In a rare joint statement, the cyber agencies of the US, UK, Australia, Canada and New Zealand warn that frontier AI models are transforming both attack and defense faster than industry expects — and urge leaders to act immediately, organization-wide and society-wide.
5
Nations issuing a rare joint cyber statement
6
Cyber agencies signing on (CISA, NSA, NCSC, ACSC, CCCS, NCSC NZ)
6
Priority defensive actions recommended now
The Core Warning
Frontier AI is a double-edged sword — lowering the barrier for attackers while becoming a powerful defensive tool.
OFFENSE ⚠
Lower barrier to entry · faster, larger-scale, more sophisticated attacks · shrinking time from discovery to exploitation
DEFENSE ✓
Vulnerability detection · anomaly monitoring · rapid response — AI as an effective shield
Six Recommended Actions — Shore Up the Basics
The message resonating most: urgently strengthen fundamentals rather than pursue exotic measures.
1
Reduce the attack surface Limit unnecessary access and external connections
2
Accelerate patching Counter shrinking time from discovery to exploitation
3
Address legacy / unsupported systems Remediate vulnerabilities in aging infrastructure
4
Strengthen identity & access management Tighten authentication and privileges
5
Prepare incident response in advance Drills and recovery plans assuming a breach
6
Use AI for defense Detection, anomaly monitoring, rapid response
Underpinned by: "secure-by-design and secure-by-default" principles + "defence in depth."
SUPPORT ✓
Growing agreement on the urgency — cyber resilience is now a board-level issue. The "defend with AI" call draws positive reactions.
CONCERN ⚠
Execution is hard in environments heavy with legacy systems and patching delays — the real test is how fast they can build defenses.
Continue reading The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.
Already purchased? Sign in ✓ Signed in — this article isn’t included in your current plan.Unlocking the full article…