Society would probably actually benefit from a non political purely objective science-based commission to review published data, make recommendations for new studies, and come up with an evidence-based recommendation to governments about whether virtual CSAM (no actual children harmed or in AI training data) and lifelike child sex dolls result in statistically more child predation.

I haven't deep dived on this so maybe it's already well known among sociologists/psych pathologists. But the key is a trusted science-based policy. We did it for violent video games and found no correlation. Not at all obvious to me if that also holds for pederasty.

Yeah I know, the trusted scientific commission is not going to happen

The case for banning simulated CSAM produced with GenAI is that if the training data contains actual CSAM then it would be directly contributing to real children being hurt. Obviously generating those pictures doesn't further cause physical harm to anyone but someone has to already have been harmed in the past for that training data to exist in the first place.

This however is not true with cartoons for example nor does it apply to sex dolls either.