imo, the danger of real images is that it’s a specific person who could be harmed if they were exposed to or made aware of content that used their face. That’s not an issue with AI-generated faces, so I don’t think there’s really any strong argument against it. If it makes you uncomfortable, don’t engage with it, but bans on any type of speech should be based on direct and proven harm. Of course, that won’t stop people from making shit up in an effort to get legal and harmless content banned, as we saw with child sex dolls.
I do think it will be regulated anyway. The US government has already indicated an interest in banning fictional texts they don’t like so I’m sure they’d be very fast to ban AI-generated images they don’t like.