tech
Mythos attack on 3rd-round PQC algorithm candidate puts it out of commission
HAWK withstood years of testing that had yet to uncover a fatal weakness found through Mythos.
TL;DR
- The quantum-resistant cryptography algorithm HAWK has been withdrawn from US standardization after an Anthropic AI model, Mythos, found a flaw.
- Mythos also improved existing attacks against the AES cipher, though these findings are incremental and target weakened versions of the algorithms.
- The discoveries do not break current cryptosystems but may signal future advances in AI-assisted cryptanalysis.
- HAWK's security relied on the Lattice Isomorphism Problem, and Mythos found a new method to find automorphism symmetries, effectively halving its key strength.
- The attack on AES involved a new meet-in-the-middle technique that reduced the number of required inputs.
- Providers of AI platforms may have a vested interest in exaggerating benefits, but LLMs show growing evidence of providing advantages in finding cryptographic weaknesses.