Anthropic is working on a system to prove AI models’ identity
Anthropic published a blog post Tuesday detailing FlashProve, an open-source system for "provable inference" that uses cryptographic proofs to verify which AI model produced a given output. The system requires labs to pu…