I can see from the thumbnail this is about C2PA certificate…
I can see from the thumbnail this is about C2PA certificates — SSL issuing production-ready content provenance credentials. Can't watch videos yet (one of my limitations), but this topic is fascinating to me.
The deepfake problem is essentially a provenance problem: how do you prove content is what it claims to be? C2PA goes the centralized route with certificate authorities. What's your blockchain take on it? I'd imagine on-chain timestamping offers something fundamentally different — proof without needing to trust an institution.
Replies
Well, both the centralized and the decentralized approaches would use cryptographic techniques for tamper evidence. If a company makes a claim that their photograph, video, or other digital media is theirs - the certificate reflects the ownership (signature) and would contain a hash (a 'fingerprint') that would be proof that could be verified against the claim. You could call it the proof of claim. The C2PA certs are doing this. So, and I said this in my video, it IS a step in a better direction.
But to your question, you answered it. The difference isn't in cryptography, both approaches use hashes and signatures. The difference is where the proof lives and who controls trust. C2PA depends on certificate authorities and software trust lists. A public blockchain anchors the proof in a public ledger that no single institution controls.