Deepfakes in real time have become famous because they allow a person to impersonate a celebrity by making him say what he wants.
but it also allows a person to say anything hiding behind an ordinary person, a sort of “videoalias”.
to know who says what, it is no longer enough to simply watch a video while listening to the voice of the speaker. seeing is no longer believing.
like today we don’t trust a piece of news written in a twit but we look at who writes it, soon we won’t trust what we see anymore, the context will be relevant for the message; we will have to know who produced it and who proposes it to us. for videos, the publisher will increase in importance.
for applications where you need to know that a specific person is who they say they are, you will need to tie the content to the process that produced it.
this may lead to integration in browsers of a software component that allows a person to authenticate with a certified digital identity (like the eIDAS scheme), enriching video metadata in a way similar to digital signature.