Have you ever thought about what would happen if AI starts making decisions for us in critical fields like healthcare and military, and it secretly has its own "agenda"?
But @Lagrange Official 's DeepProve has emerged, using zk technology to create a "behavior notarization": no matter how complex the AI is, it can prove that it uses a predetermined model and that the results are not off track.
What I look forward to most are its applications in reality: in the future, when hospitals use AI to analyze CT scans, not only can they provide diagnoses, but they can also include a "credibility certificate"; in the NFT space, those AI-generated avatars can be directly verified for authenticity.
Instead of constantly worrying about whether AI will rebel, it’s better to practically equip it with a "verifiable" brake. @Lagrange Official 's move is quite pragmatic.
