Investigation
Using Claude Mythos Preview, Anthropic researchers describe improved ways to attack cryptographic algorithms. The first result is an improved attack on HAWK, a post-quantum digital signature scheme that is a NIST Additional Digital Signatures candidate. The second identifies a new way to attack round-reduced AES. Anthropic’s own summary is careful: these are substantial research advances, but they do not currently affect any production systems.
“These are substantial research advances, but they do not currently affect any production systems.”
HAWK: ~60 hours after two years of human review
Despite HAWK having survived two rounds of expert human review over about two years, Mythos improved the best-known attack in roughly 60 hours of work — effectively cutting its key strength in half, in Anthropic’s framing. HAWK is only a candidate scheme and is not deployed; Anthropic further clarifies that neither result has a practical impact on today’s computer systems and that no production software will have to change as a result.
On AES, Mythos improved an attack on a weaker, round-reduced variant used in academic cryptanalysis — not the full cipher. Anthropic points readers to technical papers and demo code for verification, and to related evaluation work such as exploit evals for how the lab measures offensive capability.
Anthropic sources
- Discovering cryptographic weaknesses with Claude — Primary research post
- Exploit evals — Related Anthropic evaluation work