The problem is empirical data cannot be morally or ethically found. You can’t show a bunch of people porn and then make a statistical observation of whether those shown child porn are more likely to assault children. So we have to go forward without that data.

I will anecdotally observe anal sex, oral sex, and facials have gone up between partners as prevalence in porn has gone up. That suggests but does not prove a direct statistical harm caused by even “ethically produced CSAM.”

CSAM possession is illegal because possession directly supports creation

To expound on this: prior to this point, the creation of CSAM requires that children be sexually exploited. You could not have CSAM without children being harmed. But what about when no direct harms have occurred? Is lolicon hentai ‘obscene’? Well, according to the law and case law, yes, but it’s not usually enforced. If we agree that drawings of children engaged in sexual acts aren’t causing direct harm–that is, children are not being sexually abused in order to create the drawings–then how much different is a computer-generated image that isn’t based off any specific person or event? It seem to me that, whether or not a pedophile might decide that they eventually want more than LLM-generated images is not relevant. Treating a future possibility as a foregone conclusion is exactly the rationale behind Reefer Madness and the idea of ‘gateway’ drugs.

Allow me to float a second possibility that will certainly be less popular.

Start with two premises: first, pedophilia is a characteristic that appears to be an orientation. That is, a true pedophile–a person exclusively sexually attracted to pre-pubescent children–does not choose to be a pedophile, any more than a person chooses to be gay. (My understanding is that very few pedophiles are exclusively pedophilic though, and that many child molesters are opportunistic sexual predators rather than being pedophiles.) Secondly, the rates of sexual assault appear to have decreased as pornography availability has increased. So the question I would have is, would wide availability of LLM-generated CSAM–CSAM that didn’t cause any real, direct harm to children–actually decrease rates of child sexual assault?

Downloading and possession of CSAM seems to be a common first step in a person initiating communication with a minor with the intent to meet up and abuse them.

But this is like the arguments used to say that weed is a “gateway drug” by talking about how people strung out on harder drugs almost always have done weed as well, ignoring everyone who uses only weed. But this is even hazier because we literally have no real idea how many people consume that stuff but don’t ‘escalate’.

I remember reading once in some research out of Japan that child molesters consume less porn overall than the average citizen, which seems counter-intuitive, but may not be, if you consider the possibility that maybe it (in this case, they were talking primarily about manga with anime-style drawings of kids in sexual situations) is actually curbing the incidence of the ‘real thing’, since the ones actually touching kids in the real world are reading those mangas less.

I’m also reminded of people talking about sex dolls that look like kids, and if that’s a possible ‘solution’ for pedophiles, or if it would ‘egg on’ actual molestation.

I think I lean on the side of ‘satiation’, from the limited bits of idle research I’ve done here and there. And if that IS in fact the case, then regardless of if it grosses me out, I can’t in good conscience oppose something that actually reduces the number of children who actually get abused, you know?

but there’s a legitimate question as to whether we as a society decide it’s associated closely enough with real world harms that it should be banned.

Why should that be a question at all? If it causes harm, ban it. If not, don’t. Being “associated with” should never be grounds for a legal statute.

generally a very good point, however i feel it’s important to point out some important context here:

the pedophiles you’re talking about in your comment are almost always members of tight knit communities that share CSAM, organize distribution, share sources, and most importantly, indulge their fantasies/desires together.

i would think that the correlation that leads to molestation is not primarily driven by the CSAM itself, but rather the community around it.

we clearly see this happening in other similarly structured and similarly isolated communities: nazis, incels, mass shooters, religious fanatics, etc.

the common factor in radicalization and development of extreme views in all these groups is always isolation and the community they end up joining as a result, forming a sort of parallel society with it’s own rules and ideals, separate from general society. over time people in these parallel societies get used to seeing the world in a way that aligns with the ideals of the group.

nazis start to see anyone not part of their group as enemies, incels start to see “females” instead of women, religious fanatics see sinners…and pedophiles see objects that exist solely for their gratification instead of kids…

I don’t see why molesters should be any different in this aspect, and would therefore argue that it’s the communal aspect that should probably be the target of the law, i.e.: distribution and organization (forums, chatrooms, etc.)

the harder it is for them to organize, the less likely these groups are to produce predators that cause real harm!

if on top of that there is a legally available outlet where they can indulge themselves in a safe manner without harming anyone, I’d expect rates of child molestation to drop significantly, because, again, there’s precedence from similar situations (overdoses in drug addicts, for example)

i think it is a potentially fatal mistake to think of pedophiles as “special” cases, rather than just another group of outcasts, because in nearly all cases of such pariahs the solutions that prove to work best in the real world are the ones that make these groups feel less like outcasts, which limits avenues of radicalization.

i thought these parallels are something worth pointing out.