Anthropic unveils open-source tool to check which AI model replied
Anthropic released FlashProve as open source, a prototype using cryptographic proofs to verify which AI model generated a response, without revealing model weights. Reported verification overhead is 70 to 95 times.
By Standard Topic. AI-generated from the reporting listed under Coverage. Published . Updated .
This story is no longer being updated. It reflects reporting up to .
What happened
Anthropic has released FlashProve, an open-source research prototype that uses cryptography to verify whether an AI model's response came from the model a user selected. The system works without exposing the model's weights.
Timeline
- Nov 3, 4:05 AM ET: Anthropic's open-source FlashProve was reported to use zero-knowledge proofs to verify large AI model outputs without revealing weights.
- Nov 4, 10:23 AM ET: Anthropic released FlashProve as a research prototype, with significant limitations still acknowledged.
Coverage
- Anthropic Releases FlashProve To Cryptographically Verify AI Model Behind Each Response (NewsCord)
- Anthropic's new prototype proves which model wrote an AI response without exposing its weights (the-decoder.com)
More from Technology & AI
- China puts humanoid robots to work serving food and sorting parcels
- Newsom proposes sweeping AI safety rules for Washington
- Google releases Gemini 4 Neon AI model with million-token context
- Microsoft and Intel widen their cloud chip partnership
- Google DeepMind rolls out Genie 4 to select subscribers
- AI helped AMD shorten chip design work, OpenAI says