Anthropic is working on a system to prove AI models’ identity
Reported by Blockchain News, TechCrunch and Mezha3 sources ↓
Latest in this story: Anthropic Releases FlashProve To Verify AI Model Responses
Anthropic announced FlashProve, an open-source system intended to let users cryptographically verify which AI model produced an output. The research is ongoing, and Anthropic has no plans to deploy it.
Key points
- FlashProve would match outputs to model fingerprints published by AI labs
- The fingerprints are designed not to reveal a model’s weights
- Anthropic released a technical specification and working code under an open-source license
- The system is not planned for deployment at Anthropic or another lab
Sources
Details
TechCrunch reports that FlashProve is designed to provide a cryptographic proof linking an AI output to a specific model. The proposed checks are intended to work even if the model provider itself tries to substitute a different model, such as to reduce operating costs or misrepresent performance.
AI labs would publish a fingerprint for each model, which could then be matched to outputs. Anthropic’s researchers say the fingerprints would not expose model weights and that verification could run on a local device.
Anthropic released a technical specification and code for parts of the system under an open-source license. The research is still in progress, and the company said it does not plan to deploy FlashProve at Anthropic or elsewhere.
Written from coverage as of
Images
- TechCrunch
Stocks in this story
Change since open: 20 Nov to 21 Nov 2026
Last refresh