Tech
30/07/2026Résumé IA
Anthropic's AI Uncovers Vulnerabilities in Cryptographic Algorithms, Including US Federal Standard Contender HAWK
Anthropic's AI model, Claude, discovered two previously unknown attacks on cryptographic algorithms: one against HAWK, a post-quantum digital signature scheme competing for US federal standard status, and another against a research version of AES. The HAWK vulnerability significantly reduces the security margin, potentially impacting its competitiveness due to required key size increases, while the AES attack demonstrates a novel approach though against a non-standard configuration.
D
Decrypt5 min de lecture
