A post-quantum cryptographic algorithm named HAWK has been withdrawn from consideration as an official US standard following the discovery of a significant weakness by Anthropic’s Mythos AI security model. HAWK is a digital signature scheme designed to resist attacks from quantum computers and had advanced through two rounds of testing by the National Institute of Standards and Technology before encountering the vulnerability during its third evaluation round.
Anthropologic researchers used approximately 60 hours of computational time and roughly $100,000 in computing resources to have Mythos identify a previously unknown method for attacking HAWK’s underlying mathematical foundation. The attack targets the Lattice Isomorphism Problem, which HAWK’s security depends upon, and effectively reduces the algorithm’s key strength by half. The weakness could theoretically be addressed by doubling key sizes, but this additional computational burden makes HAWK less competitive compared to alternative post-quantum signature schemes already in development. Experts in cryptography have characterized the algorithm as no longer viable following this discovery.
Anthropologic also reported using Mythos to identify improvements against AES, a widely used encryption cipher, though these results were less dramatic. The company demonstrated that the AI model could reduce the number of plaintext inputs required for a theoretical attack from approximately 2^105 to 2^89, representing a substantial computational savings. However, researchers emphasized that such attacks remain impractical outside controlled laboratory environments.
Experts caution that while these findings are noteworthy, they involve weakened versions of cryptographic systems rather than production-standard implementations, and the underlying mathematical foundations remain secure for the present. The research highlights the potential for artificial intelligence to contribute to cryptanalysis while raising questions about whether AI systems can be leveraged to discover cryptographic weaknesses that traditional methods might eventually uncover independently. The broader implications for cybersecurity remain uncertain as this emerging capability continues to develop.
Article Attribution | Read More at Article Source
Article summary produced by Claude AI