This is a tough one, it certainly makes me very uncomfortable, but if I really think about it, I'm not sure it's gonna have the detrimental impact that some people think it will. (is there a way to make this thing stop trying to guess what I'm about to say? It's driving me nuts) I mean the way I see it, no matter how good it gets, it will not be able to convincingly mimick the most unique people out there. Like I highly doubt it would ever be able to make a convincing Markiplier video, or even a Ted Nivison etc, because these people surprise us, they have a style but the style is spontaneous, they do weird voices and impressions and scream etc. I don't even think 10 years from now AI will be able to convincingly replicate THAT.
I DO think it will (and maybe already can) convincingly be able to replicate lower energy "NPC" types, people who talk the same way every video, just deliver information, the more monotone among us etc. But here's the thing, if that's you? Why would someone want to use AI to steal that anyway? That's not unique enough for someone to WANT to steal. (Also, if you have that kind of delivery, I'm not saying that's bad or anything, I don't mean "NPC" negatively, a lot of people do research and do awesome informative videos with a monotone delivery). So I guess I"m just not sure who's really gonna suffer from this, if you try to trick us into thinking it's a truly animated unique person we all know? It probably will never be able to get the nuance right. If it tries to mimick the monotone? First off why bother what's the point, and secondly we'd probably STILL know because the type of info that person usually delivers will feel off you know?
I'm sure it will be able to fool some, I just doubt it's gonna be able to fool everyone, and these things will probably get found out eventually. What seems more likely to me is them making short clips that make it look like someone is saying something offensive and people flying into a rage before vetting the video. THAT does worry me.