it functioned just fine for the past 50 years. digital manipulation of media had always been possible. americans simply didn't have a society-wide psychosis during the early days of personal computing and Internet.
"make the person in this photo naked. No mistakes"
Prompts lowered the bar. Even if you do have to install a local text to image model.
there are, for example, websites that host (among a myriad other such things) style-accurate pictures of every Disney princess being fucked in every way imaginable. 70 years ago something like this might have caused a similar moral panic, but now no one (rightfully) gives a shit. you see a picture of Cinderella with a dick up her ass, you know it wasn't drawn by Disney.
Theyre not the same thing and you can't just hand wave away people's pain. You're part of the problem.
Since perfect or not people aren't going to stop doing what they do, the fact that the fakes have gotten better is only beneficial since it allows for plausible deniability. At this point nobody should trust some random video the same way nobody has been able to trust a random JPG since photoshop became a thing.
Yes, from a victim impact perspective.
> Before computers people would cut out peoples faces and paste them onto playboy centerfolds. That sure wasn't perfect, but it didn't stop them
The effort did stop them. The problem wasn't even half as widespread. And those who were doing it were keeping it to themselves, the images weren't being auto-shared to Twitter and the targets tagged.
> the fact that the fakes have gotten better is only beneficial
[Victim testimony needed]
> At this point nobody should trust some random video
While there's a qualitative difference in harassment when using media that is either real or fake (clearly or not) the capacity for that media to be used for harassment purposes remains the same.