[y4mi]

It's called agree of misinformation because it's become increasingly easy to produce convincing fake footage.

It's entirely possible to fake a real-time video interview for instance using voice synthesis and mapping the face of your target upon your face. It's frighteningly convincing.

I.e. https://m.youtube.com/watch?v=AmUC4m6w1wo

[eeeuo]

I don't think that footage is frighteningly convincing.

It's pretty good, but not difficult to tell that it's not real. The bridge between pretty good and perfect when doing this type of thing is a very, very difficult problem.

It also requires substantial amounts of high quality, up close footage of the person. This may be work for major politicans, but is much more difficult for the vast majority of people.

I don't disagree that it will become more of a problem in the future, and in many cases "pretty good" will be sufficient, but I think people are overstating how much of a problem it will be.

[y4mi]

It's frightening because this is a the work if students with consumer hardware.

Video material can be produced at will if a wealthy nation or corporatiom wants it and the model will be significantly better if they train it on specialized hardware.

I don't think that the average person has to worry about his face being stolen. I think that it's possible to produce extremely good footage if somebody is sufficiently motivated to fund a smear campaign. It won't hold up in court, but it doesn't have to either.