>Well in my case I specifically mean hyper realistic deep fakes of people created without their consent.
Seems reasonable on its face, but the real issue then becomes damage done-vs-effort needed to prevent it
Your example is murder, but the problem is then a scaling issue; not many people really want to commit murder. It takes effort, and the sample size is so small we can evaluate it on a case-by-case basis. But it's trivial to deepfake one face onto another, so it's likely a lot of people will try it for their own amusement.
So basically murder to most people is brutal and ugly. A Deepfake is clicking a few buttons; it's abstract and hard to see as a real crime for most people.(...ya know, kinda like how I'll pirate a song but I'm not going to stuff CDs into my pockets at the local retail store)
And now we can expand that idea- what counts as a deepfake? The face editing is obvious here but what about training an AI to replicate someone else's voice? What about using the Lamba character.ai site to replicate Sonic or Indiana Jones? That's where it really gets tricky and where I feel politicians will be at a loss to construct any meaningful legislation.