Claude Discovered Flaws in Cryptographic Algorithms Missed by Years of Expert Review
Claude found mathematical flaws in two cryptographic algorithms that years of expert review missed.
News
-
Events
- TNW Conference: June 19 & 20, 2025
-
Newsletters
-
Partner with us
-
Jobs
-
Contact
Latest
-
Deep tech
-
Sustainability
-
Ecosystems
-
Data and security
-
Fintech and ecommerce
-
Future of work
Conference Media Hub
-
More
-
Startups and technology
-
Investors and funding
-
Government and policy
-
Corporates and innovation
-
Anthropic Announces Claude Mythos Discoveries
Claude Mythos, a powerful AI tool, has made significant breakthroughs:
-
HAWK Algorithm Flaw: Halved the key strength of HAWK, a post-quantum digital signature scheme under NIST review, in just 60 hours.
-
AES Attack Improvement: Accelerated attacks on reduced AES by 200-800x, autonomously discovering a technique named "Möbius Bridge."
Key Takeaways:
- Neither discovery affects production systems as HAWK is not deployed, and the AES attack targets a reduced variant.
- Each discovery cost approximately $100,000 in API compute.
- Anthropic followed responsible disclosure practices, sharing with authors and coordinating with relevant stakeholders before publication.
AI’s Role in Cybersecurity Evolution
Anthropic highlighted a significant shift: moving from bug identification in implementations to discovering flaws in cryptographic algorithms themselves. This represents a substantial advancement in what AI can contribute to security infrastructure.
Follow-Up Results and Implications
Anthropic also acknowledged further results, including attacks on LEA, Serpent-128, Salsa20, Poseidon, and SHA-1, emphasizing the rapid pace of AI’s evolution in cryptography.
The White House’s Gold Eagle initiative coordinates AI-powered cyber defense but lacks an equivalent program for cryptographic review, raising important questions about future developments in this domain.