The Five Eyes intelligence alliance has cautioned that the rapid advancement of artificial intelligence could allow cyber threats to evolve faster than existing security practices can keep up.
The warning comes amid growing concerns over AI-powered hacking capabilities, particularly after US-based AI company Anthropic revealed in April that its advanced Mythos models demonstrated exceptional skill in identifying software vulnerabilities.
In a joint advisory issued on Monday, the intelligence agencies of the United States, Britain, Australia, Canada, and New Zealand urged governments and businesses to strengthen their cyber resilience as AI technology continues to advance.
āThe rapid pace of frontier AI development means cyber risk assumptions can become outdated in months, not years,ā the statement said.
According to the advisory, AI ālowers barriers for malicious actors and increases the speed and complexity of attacksā.
āBreaches will occur. Preparedness helps you contain them quickly and prevent escalation into major operational and financial crises.ā
The alliance recommended that organisations enhance cybersecurity by incorporating AI into security operations, modernising outdated infrastructure, and restricting access to sensitive systems.
Separately, Anthropic suspended access to its Mythos 5 and restricted Fable 5 models this month after receiving a US national security directive. The move came shortly after the public release of Fable 5, with the government order prohibiting foreign nationals from accessing both models.
The decision is notable given the White House’s broader push to reduce AI regulation and limit states from introducing their own oversight measures.
