The Five Eyes intelligence alliance—comprising the United States, United Kingdom, Canada, Australia, and New Zealand—has issued a stark warning that artificial intelligence models capable of launching devastating cyberattacks against governments and businesses are only months, not years, from becoming operational. In a rare joint statement, the alliance urged immediate action to strengthen cyber defenses against this emerging threat.
Urgent Call to Action
The warning comes amid growing unease among Western nations about the rapid advancement of AI capabilities. The Five Eyes statement emphasized that frontier AI models are expected to exceed current industry expectations, fundamentally transforming both offensive and defensive cyber operations. “The timeline is not years, it is months,” the spy agencies declared, underscoring the urgency of the situation.
Anthropic Restrictions and Jailbreaking Concerns
The alert follows the Trump administration’s unprecedented directive ordering AI company Anthropic to suspend use of its most advanced models—Mythos 5 and Fable 5—by foreign nationals. This marked one of the farthest-reaching government actions in response to AI model capabilities. Anthropic’s Mythos model had raised widespread cybersecurity concerns due to its exceptional ability to identify security flaws. The company believes the US government discovered a method to jailbreak its public Fable model, bypassing internal safety guardrails. Anthropic and the administration are meeting to resolve the issue.
Lowering Barriers for Malicious Actors
AI researchers and executives have long expressed safety concerns about advancing technology. The Five Eyes leaders noted that AI lowers barriers for malicious actors while increasing the speed and complexity of attacks. Cybersecurity experts describe the message as stark, with worrying implications not only for governments and large corporations but also for small and medium-sized businesses worldwide.
Recommendations for Defense
To counter the threat, the Five Eyes alliance advised businesses and leaders to invest in cyber defenses, upgrade outdated systems, patch faulty software, and limit access to critical systems. They also highlighted that AI itself can be part of the solution. “Organizations that integrate AI tools into their security operations can detect vulnerabilities earlier, improve software quality, monitor unusual behavior, and respond faster to incidents,” the security alliance stated.
Open Letter on AI Risk Assessment
Dozens of cybersecurity researchers, AI entrepreneurs, and corporate executives recently signed an open letter urging the Trump administration to commit to an open, scientific, and transparent process for handling AI risk assessments. The letter emphasized that such transparency is essential for security teams to find and fix flaws in both newly written code and decades of legacy code faster than adversaries.
