
Artificial Intelligence
Anthropic’s FlashProve prototype verifies AI responses without revealing model weights
Anthropic has released FlashProve, an open-source research prototype that uses cryptography to verify whether an AI response was generated by a claimed model. The system is faster than earlier comparable tools, but its current limits make it unsuitable for many everyday chatbot exchanges.


