Anthropic says Claude uncovers cryptographic weaknesses, advances AI-assisted security research

The company noted that while modern cryptographic systems remain highly secure, AI models are increasingly capable of assisting researchers in analyzing complex cryptographic constructions