Five Eyes warns AI could outpace cyber defenses within months
Why it matters: The alliance urged governments and companies to harden systems now as AI speeds up and lowers the barrier for cyberattacks.
The Five Eyes intelligence alliance warned that fast-improving AI models could outstrip current cybersecurity assumptions within months, raising the risk of more complex and faster attacks. In a joint advisory dated Monday, agencies from the United States, Britain, Australia, Canada and New Zealand said AI can help malicious actors find and exploit software weaknesses more easily. The warning follows Anthropic's April disclosure that its Mythos models showed unusual skill at identifying vulnerabilities. The alliance urged organizations to add AI tools to security operations, modernize older systems and tighten access to critical networks so breaches can be contained before they spiral into operational or financial crises.