However, I think it's clear that Pandora's box is now wide open and that you cannot close it. Sure, you can turn off that Grok integration, but the AI image generation capabilities are now widely available for basically anyone to use.
I wonder whether it'd be better to just "accept and live with it". I agree that this can cause a lot of harm, but I don't see a way where this can be outlawed and prosecuted in such a way that there's a net benefit for society. In the EU many have been battling proposals like Chat Control for the past decades not because they want to protect sex offenders, but because backdooring societies privacy on a grand scale is likely far more detrimental than the impact of sex offenders. (And here we aren't even talking about "real" CSAM content.)