CG

How does verifiable AI execution work?

Updated May 29, 2026, 12:19 PM

It employs zk-proofs (such as PLONK) and TEE attestations to cryptographically prove model execution correctness without exposing sensitive data.

View original source