[Anthropic is developing an open-source system called FlashProve to cryptographically prove the origin of AI output] According to IT Home, Anthropic is developing an open-source system called FlashProve to verify, through cryptographic proof, which model generated a given AI output. The system aims to achieve "provable reasoning," providing users with mathematical guarantees that the received answer indeed comes from the model they believe they are using, and preventing the impersonation of one model as another. FlashProve works similarly to encrypted email signatures, providing a "fingerprint" for each model that does not reveal its weights, which is then matched against a specific output for verification. Researchers say the system has the potential to prevent AI companies from secretly switching to cheaper models, or companies from using more expensive models to inflate test results, and can also verify third-party claims that a task was performed by a more advanced model. The system is lightweight, allowing users to perform verification on their local devices.