Depending on which way it goes, it could be massively helpful for protecting kids

Weeeelll, only until the AI model needs more training material…

In Australia cartoon child porn is enforced in the same way as actual child porn. Not that it answers your question but it’s interesting.

I’d imagine for your question “it depends”, some people who would have acted on their urges may get their jollies from AI child porn, others who have never considered being pedophiles might find the AI child porn (assuming legal) and realise it’s something they were into.

I guess it may lower the production of real child porn which feels like a good thing. I’d hazard a guess that there are way more child porn viewers than child abusers.

Do we know that AI child porn is bad? I could believe it would get them in the mood for the real thing and make them do it more, and I could believe it would make them go “ok, itch scratched”, and tank the demand for the real stuff.

From bits/articles I’ve seen here and there over the years about other things that are kind of in the same category (porn comics with child characters in them, child-shaped sex dolls), the latter seems to be more the case.

I’m reminded of when people were arguing that when Internet porn became widespread, the incidence of rape would go through the roof. And then literally the opposite happened. So…that pushes me toward hypothesizing that the latter is more likely to be the case, as well.

Real question: “do we care if AI child porn is bad?” Based on most countries’ laws, no.

I’d like to know what psychologists think about it. My assumption is the former, it escalates their fantasizing about it and makes them more likely to attack a child.

There seems to be no way to conduct that experiment ethically, though.

You’re missing the point. They don’t care what’s more or less effective for helping kids. They want to punish people who are different. In this case nobody is really going to step up to defend the guy for obvious reasons. But the motivating concept is the same for conservatives.

I seem to remember Sweden did a study on this, but I don’t really want to google around to find it for you. Good luck!

There definitively is opportunity in controlled treatment. But I believe outside of that there are too many unknowns.

In Canada even animated cp is treated as the real deal

There’s like a lot of layers to it.

• For some, it might actually work in the opposite direction, especially if paried with the wrong kind of community around it. I used to moderate anime communities, the amount of loli fans wanting to lower the age of consent to 12 or even lower was way too high, but they only called people opposed to loli as “real predators”, because they liked their middle-school tier arguments (which just further polarized the fandom when the culture wars started).
• Even worse might be the more realistic depictions might actually work against that goal, while with (most) loli stuff, at least it’s obvious it’s drawn.
• An often overseen issue is, data laundering. Just call your real CP AI generated, or add some GAI artifacts to your collection. Hungary bans too realistic drawings and paintings of that kind, because people even did that with traditional means, by creating as realistic tracings as possible (the calling CP “artistic nudes” didn’t work out here at least).

Wikipedia seems to suggest research is inconclusive whether consuming CSAM increases the likelihood of committing abuse.