Sure it's distasteful, but it's no different to cutting and pasting heads onto porn stars as was done plenty before.
I'm not saying there's entirely no harm in that, it's obviously a form of bullying, but AI does not make it novel, or a crisis.
Sure it's distasteful, but it's no different to cutting and pasting heads onto porn stars as was done plenty before.
I'm not saying there's entirely no harm in that, it's obviously a form of bullying, but AI does not make it novel, or a crisis.
(Alas I guess that might be true for this lesson too!)
Why would we assume your scenario will occur with even more effective bullying if it isn't already with that?
"hey kids, get used to being exploited sexually, as it would be too expensive to require massive multinational corporations to bother to regulate AI"
Open models cannot be regulated. The genie is out of the bottle permanently.
That depends on what they're exposed to, how much of it, and how quickly.
Too much nastiness and you get PTSD and catatonia, like kids in Gaza or child abuse victims. Or suicide.
Children need to experience bad stimuli, for sure. But that doesn't mean they can cope with unlimited really bad stimuli.
> Do you think it comes from hearing repeated "It's OK" after a child fell?
No. But it doesn't come from "you fucking idiot child, you can't even walk properly, you'll never amount to anything" either.
It is bullying.
But AI does not make it more than that, or a crisis.
edit: Oh I get it now, we have some lobby group posting here.
A bullying machine that makes this sort of extreme bullying trivial and leads directly to widespread use of the bullying machine is a massive problem.
Furthermore, most people are not technical in nature and cannot tell the difference between deepfakes and real photos and videos in short bursts.
Basically, the friction needed to develop revenge porn has been dramatically reduced.
They don't need to. The point is that eventually, everyone will just assume it's a fake, exactly because you can't tell, and fakes are easier to produce and thus more common than real leaked nude photos and videos. At that point, a deepfake shouldn't be more socially damaging than a rumor.
Rumors can be exceedingly damaging - to the point of death - even without convincing photographic evidence of it.
But the fancier tooling changed things a bit each time.
After social media became common it was also hypothesized that embarrassing stories dug up from somebody’s last would not be harmful anymore but that future has never materialized.
this is not a good, nor should it ever be, an inevitable thing
That said, there's the question of where the line is drawn. If I clip from a biology textbook an image of a nude woman, and I clip the Gerber baby's head and paste it onto the nude woman's head, what is that? If I generate the face of a young boy, then draw a stick figure under it and give the stick figure a stick erection, what is that? I don't know. I mean these things are weird and offensive, but to what degree?
There were a few really big stinks about this too, with the school director getting involved.
I mean, this is not new. Just evolved.
You must not have kids. If you did, I’m speculating that you’d get it.
If you do have kids (or if you can empathize having kids), would you be ok with tech that super easily allows your kids’ peers to share/laugh about nudes of your kids?
You can kill with a gun, a knife, a fork. Removing the tool does not change the situation of being around a person who wants to kill you.
We homeschool our kids and they have grown up to be respectful people.
It's that AI is not novel and does not present a crisis.