> I am aware of the purported benefits of decriminalizing certain kinds of CSAM, but utilitarian arguments are suspiciously silent on the subject of costs.

The reason they're suspiciously silent is because there is little to no evidence for the existence of such costs. The harms of CSAM with real children is clear as the victim is on camera, but who is victimized by animated CSAM? The allegation is that exposure to porn encourages deviant thoughts that eventually culminate in molestation. This is contradicted by lots of sex research in ordinary people and porn, and the small studies of child molesters suggests that it reduces recidivism too, like I said.

So the situation we're in is that the government has banned something with little to no evidence of actual harm. It's the obscenity laws against porn all over again. Shouldn't the government ideally have to demonstrate actual harm before regulating behaviours?

I understand the impulse to the precautionary principle, but it doesn't seem warranted here.