BlockBeats News, July 29th, Anthropic announced that researchers have successfully discovered two improved cryptographic attack methods using the Claude Mythos Preview model. The first attack significantly weakened HAWK, a digital signature scheme designed for the post-quantum era; the second identified a new attack path against the simplified round AES, which is currently the most widely used symmetric encryption standard.

Both discoveries represent significant research advancements but currently do not impact any practical production systems. Anthropic sees this as a crucial validation of the potential of AI models in the field of cryptographic security research.

Click on the original article link below to join the BlockBeats · Feishu AI News Channel, monitoring global AI hot topics and news 24/7.