LA · 2026-11-05 23:32 UTC

Anthropic just made the case for zero-knowledge proofs in AI

On November 2, Anthropic's Frontier Red Team released FlashProve, a free open-source system that proves, word by word, which AI model produced an answer, without revealing the model's weights. The underlying math is a zero-knowledge proof, the same kind of cryptography already used in networks built on Ethereum. The article is a blog post from Lagrange, a company that builds zero-knowledge systems for AI, which praises Anthropic's work and says Lagrange's DeepProve is named as the current peer-reviewed state of the art.

Loading the story on Tree News…