Anthropic Uses Claude to Detect New Cryptographic Attacks

Anthropic has reported that its Claude Mythos Preview model has identified new ways to attack the cryptographic algorithms HAWK and a simplified version of AES. Claude found a vulnerability in HAWK, reducing the effective cryptographic strength of keys by half within 60 hours of operation. HAWK is currently not used in real systems and is undergoing the standardization process. Anthropic has shared the findings with the authors of HAWK and coordinated the disclosure with NIST. Regarding AES, Claude discovered a more effective attack method on an experimental version with seven rounds, operating 200 to 800 times faster than previous methods. This finding does not threaten the full version of AES, which is used in modern products. The research cost about $100,000 on the API. Claude also identified new attacks on simplified versions of the LEA and Serpent algorithms, as well as improved analysis methods for Salsa20, Poseidon, and SHA-1. Anthropic has introduced the CryptanalysisBench benchmark to evaluate the capabilities of language models in cryptanalysis.

This content is provided for general informational purposes only and doesn't constitute financial, investment, legal, or tax advice. Any events, rewards, online promotions, or related information mentioned herein should not be considered a recommendation, solicitation, or invitation to purchase, sell, trade, or otherwise deal in any crypto assets. Crypto assets are highly volatile and may result in loss. The availability of WEEX services, products, and related events may vary by region. You are responsible for ensuring that your participation is in accordance with applicable local laws and regulations.