Comment by mschuster91

> Truly perverse incentives such as one the one you linked aside I think it's mostly fine.

Well, the thing is, with AI being widely available the threat model explodes as the difficulty goes down drastically - imagine someone deepfaking a video of a C-level executive being involved in illegal or "extreme" sexual acts; we already have "nudifier" apps, the steps to cross for the mentioned scenario aren't that large. The number of potential threat actors explodes as the group is now "everyone with a smartphone", and it also explodes as the likelihood of getting caught (and sentenced to decades in jail, if not death) for shooting someone in public is significantly higher than getting caught "leaking" a faked video which at worst risks you a year or two for defamation.