Five Eyes warn that AI capable of destabilizing governments could arrive within months

The Five Eyes intelligence alliance issued a rare joint warning that highly advanced AI models capable of disrupting governments and major businesses may be only months away. The statement follows the Trump administration’s decision to block foreign nationals from accessing Anthropic’s Fable AI model, reflecting growing national security concerns. Agencies caution that frontier AI will dramatically accelerate cyberattacks by lowering barriers for malicious actors and increasing the speed and sophistication of threats. Anthropic’s Mythos and Fable models are cited as examples of tools powerful enough to detect system vulnerabilities, prompting fears that similar or even more advanced models may be under development globally. The agencies urge governments and businesses to treat AI driven cyber risk as a core leadership issue requiring a whole of society response.

Read the full story on The Guardian →