AI Accelerates Vulnerability Research, Exposing Cryptocurrency Weaknesses
Anthropic's AI model, Claude Mythos Preview, has discovered new methods to attack cryptographic algorithms, highlighting the growing concern of AI in cryptocurrency vulnerability research. The model found a vulnerability in HAWK, a candidate for a future quantum-resistant signature standard, in just 60 hours.
A bug introduced in 2021 had reduced the entropy of certain seeds generated by several Coldcard models, and experts believe an attacker may have used a Large Language Model (LLM) to generate scripts that automated these thefts.
Bitcoin Core developers are aware of this evolution, with more security reports and contributors using AI. However, they emphasize that AI primarily acts as a multiplier, allowing competent developers to review more code and test more scenarios.