Anthropic’s Claude Mythos Uncovers Flaws in Post-Quantum Cryptography

Anthropic’s Claude Mythos AI has revealed new vulnerabilities in the HAWK post-quantum signature scheme, demonstrating AI’s growing impact on cryptographic security.

Anthropic’s AI model, Claude Mythos, has identified vulnerabilities in critical cryptographic algorithms, including an improved attack on HAWK, a post-quantum signature scheme. According to The Decoder, this scheme had undergone human expert review for more than two years before the AI found its weaknesses.

The AI model discovered these vulnerabilities in just 60 hours, incurring an API cost of approximately $100,000, The Decoder reported. Although these findings do not impact currently deployed systems, they highlight the potential for AI to challenge fundamental assumptions underlying internet security.

For Japanese investors and market participants, this development signals the increasing role of AI in cybersecurity and cryptography, emphasizing the need to monitor how emerging technologies could affect digital asset security and financial infrastructure.

#newsroom

By Mansa Kane

Tokyo Overnight

Japan markets, decoded. Get the brief in your inbox.

No spam. Unsubscribe anytime.