It has nothing to do with deepfakes. The problem is that posts on any social media (and most chat applications) have no signatures, it's impossible to know that they're actually written by the user that posted it. This problem has been solved 50 years ago, social media platforms are literally 50 years behind the times; does that make sense now?

I can verify, with and without a computer, now and till the end of time, that "5885893a033d01aab4d337375a740c043a8a6241e868873084342a658c3f924c1e43a140111025fd8b5f6a8135f1d077c9fd6c80958834681615bb60be12069e" is a valid signature by "9aff936f8619411c0692e14d7aaf49e45cfd908e8a9c7afd3c72a64ba0dcabfc" for note 3002d0906592f1f372bfedfc0c61c5e770c2a37e7c84d4f267edc8995f544229 (the one I'm replying to), meaning that the owner of that npub authored (you) authored that post. I don't have to rely on a central authority for that, a relay owner CANNOT submit a note and pretend you sent it, no matter how many resources they have (let's ignore quantum computers, once they become actually real (and I mean actually usable) nostr can easily switch to a quantum resistent algorithm and apps can display warnings over non-quantum signatures).