Cyber agencies in the Five Eyes alliance have urged businesses and organizations to prepare, as frontier AI models capable of enabling major cyberattacks could emerge within months.
“As the leaders of the Five Eyes cyber security agencies, we are united in our call to action: the evolving landscape of artificial intelligence (AI) is rapidly transforming cyber risk, and we must act swiftly to remain ahead,” the Five Eyes alliance, comprising the United States, Australia, the United Kingdom, Canada and New Zealand, said in a joint statement.
“Frontier Al models are anticipated to exceed current industry expectations, fundamentally transforming both offensive and defensive cyber capabilities. The timeline is not years, it is months,” the statement added.
The intelligence agencies said organizations should prepare for the possibility of cyber breaches and strengthen defenses by limiting unnecessary access to systems, promptly applying security updates, testing incident-response plans, strengthening identity authentication measures and restricting access to critical infrastructure.
“Breaches will occur. Preparedness helps you contain them quickly and prevent escalation into major operational and financial crises,” the statement said.
The alliance also encouraged businesses and government agencies to incorporate AI tools into their own cybersecurity operations.
“Adversaries are already using AI to move faster and more effectively. Defenders must do the same,” the agencies said. “Organisations that integrate AI tools into their security operations can detect vulnerabilities earlier, improve software quality, monitor unusual behaviour, and respond faster to incidents – reducing both the cost and impact of incidents.”
While the statement did not name any specific company or model, it comes amid heightened scrutiny of advanced AI systems developed by Anthropic.
Earlier this month, the Trump administration blocked foreign nationals from accessing Anthropic’s advanced AI models, including Fable and Mythos, citing national security concerns.
Anthropic’s Mythos model, released earlier this year, is designed to identify vulnerabilities in cyber systems and has been made available only to vetted organizations due to concerns it could be misused. The company has claimed that the AI tool can outperform humans in certain hacking and cybersecurity tasks.
