Anthropic develops FlashProve for provable AI model identity
Original title: Anthropic is working on a system to prove AI models’ identity
An open-source system offering cryptographic proofs that outputs come from claimed models without revealing weights; lightweight and local verification possible.
Anthropic is developing FlashProve, an open-source system for provable inference using cryptographic proofs to verify AI model outputs. It resists model impersonation by labs themselves and supports verifying third-party claims like vulnerability scans or business approvals. Researchers released technical specs and code on GitHub; labs must publish model fingerprints for matching. Research is ongoing with no deployment plans.
Unconfirmed: Research status as intermediate progress only
- FlashProve is lightweight for local device verification
- Fprints match outputs cryptographically without exposing weights
- Designed to stop secret model swaps or performance inflation
- Supports verifying frontier model claims on tasks
- Open-source under permissive license
- Vendors
- Anthropic
- Products
- FlashProve
- AI models
- Anthropic
From the original article
2 min read at techcrunch.com
Called FlashProve, the system would let anyone check that an AI response really came from the model it claims to be from.
By Russell Brandom · techcrunch.com
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.