Anthropic is developing a response verification system called FlashProve

The company Anthropic has begun developing a new authentication system called FlashProve. The project aims to address the challenge of verifying the origin of AI-generated content.

The new tool will allow any user or third-party platform to technically confirm that a specific response was generated by the claimed language model. Such cryptographic verification prevents the falsification of results and the passing off of responses from third-party systems as someone else's work.

The development comes amid the growing prevalence of unlicensed and disguised models in the market. Implementing FlashProve will help developers and corporate clients guarantee the authenticity of AI agents and protect intellectual property.