I agree with you. I think using an AI tool to figure out what type of voice you want and then hiring someone to perform the actual game dialogue would be more appropriate. It would be like using ai images to help a storyboard project. I guess you could use an artist for that and have them do different iterations is an argument against it. Voice acting seems to be more limited unless you know who will be the voice despite knowing the end target. Early dirty work that never makes it further is the only place I want to see AI in gaming. Unless the game is based around AI as the core component like procedurally generated rogue likes. I don’t enjoy that style of game so I’d be fine with that not happening either.