Obviously AI generated images and videos are already being used in politics, generally as far as I’ve seen by right wingers.
A thought I had would be to flip the table and generate AI videos of Trump attacking right wingers for being too far right, too racist, not supporting trans or LGBT+ rights, etc, in ways that are guised as old school conservative values. Like the “The govt shouldn’t be interfering in a private citizen’s healthcare choices” type stuff.
His supports seem the type already to believe AI videos are real, so do you think stuff like that would work or just circle in left wing groups and never take off among his worshipers? If they did spread, would they even be effective in triggering his base to push back towards the center?
As much as I strongly (very strongly) dislike AI-generated anything, I can at least acknowledge that there can be ethical uses of it assuming you hold your nose and ignore the very likely unethical means by which the model(s) were trained.
That said, what you described would definitely not be ethical IMO. I see where you’re coming from, I do, but I’ve never been one to fight disinformation/lies with more disinformation/lies. Beyond the slippery slope, it’s still operating in bad faith and not something I could ever consider ethical let alone welcome.