[jasonmorton]

One can cryptographically prove the correct inference of a small AI model now with https://github.com/zkonduit/ezkl (our open-source package).