Anthropic says its Mythos model found vulnerabilities in cryptographic algorithms that secure the internet

Anthropic has revealed weaknesses in key cryptographic algorithms through its Claude Mythos model. It discovered vulnerabilities in the HAWK post-quantum signature scheme, which had been reviewed by human experts for over two years, in just 60 hours at a cost of approximately $100,000. The findings indicate that current systems are not affected, but they highlight how AI could challenge core assumptions about internet security. This discovery reflects the ability of intelligent models to identify vulnerabilities faster than human capabilities.