The intelligence alliance known as “Five Eyes” is warning that powerful new artificial intelligence models are rapidly becoming capable of devastating cyberattacks against governments and businesses. In a rare joint statement, intelligence officers from the U.S., U.K., Australia, New Zealand and Canada warn that generative AI models could soon lower barriers for malicious hackers, while vastly increasing the speed and complexity of cyberattacks.

This comes after the Trump administration ordered Anthropic to limit access to its Fable and Mythos AI models after they proved highly adept at exploiting critical software vulnerabilities. According to a report in The Economist, an Anthropic AI agent was able to penetrate nearly all classified systems managed by the NSA and U.S. Cyber Command within hours.