Anthropic Open-Sources 'FlashProve', a System That Verifies Which Model Produced an AI Output

IBL News | New York

Anthropic released FlashProve, an open-source prototype that cryptographically verifies which AI model produced an output, the San Francisco company announced this week.

The system provides what researchers call “provable inference” — a cryptographic proof, essentially a mathematical guarantee, similar to cryptographic email signatures.

It lets users confirm that the answer they received really came from the model they think they’re using.

FlashProve would be robust against impostors attempting to disguise one model as another. Researchers note that the system could prevent an AI company from secretly swapping in a cheaper model to cut costs.

To work, the system requires AI labs to publish a “fingerprint” for each model they release, which can later be matched to specific outputs and cryptographically verified.

IBL News | New York

Anthropic released FlashProve, an open-source prototype that cryptographically verifies which AI model produced an output, the San Francisco company announced this week.

The system provides what researchers call “provable inference” — a cryptographic proof, essentially a mathematical guarantee, similar to cryptographic email signatures.

It lets users confirm that the answer they received really came from the model they think they’re using.

To work, the system requires AI labs to publish a “fingerprint” for each model they release, which can later be matched to specific outputs and cryptographically verified.