Oh boy, the future is now:

Welcome to TrumpOrBiden2024, the ultimate AI generated debate arena where AI Donald Trump and AI Joe Biden battle it out 24/7 over topics YOU suggest in the Twitch chat!

https://www.twitch.tv/trumporbiden2024

It’s best to just view a few minutes of it yourself, but this rank absurdity is an AI-driven “deep-fake” endless debate between Trump and Biden, full of NSFW profanity, and somehow splices in topics from Twitch chat. As the Kotaku article mentions:

The things the AI will actually argue about seem to have a dream logic to them. I heard Biden exclaim that Trump didn’t know anything about Pokémon, so viewers shouldn’t trust him. Trump later informed Biden that he couldn’t possibly handle genetically modified catgirls, unlike him. “Believe me, nobody knows more about hentai than me,” Trump declared. Both men are programmed to loosely follow the conversation threads the other sets, and will do all the mannerisms you’ve come to expect out of these debates, like seeing Biden react to a jab with a small chuckle. At one point during my watch, I saw the AI stop going at each other only to start tearing into people in the chat for having bad usernames and for not asking real questions.

It’s interesting how far we have come as a society and culture. At one point, deep-fakes were a major concern. Now, between ChatGPT, Midjourney/Stable Diffusion, and basic Instagram filters, there is a sort of democratization of AI taking place. Granted, most of these tools were given out for free to demonstrate the value of the groups who wish to eventually be bought up by multinationals, but the things developed in such a short time is nevertheless amazing.

Of course, this all may well be the calm before the storm. Elon Musk’s lawyers tried to argue that video of him claiming that Teslas could drive autonomously “right now” (in 2016) were in fact deep-fakes. The judge was not amused and said Elon should then testify under oath that it wasn’t him in the video. The deep-fake claim was walked back quickly. But it is just a matter of time before someone gets arrested or sentenced based on AI-fabricated evidence and when it comes out as such, things will get wild.

Or maybe it won’t. Impersonators have been around for thousands of years, people get thrown into jail on regularly-fabricated evidence all the time, and the threat of perjury is one of those in-your-face crimes that tend to keep people honest (or quiet) on the stand.

I suppose in the meantime, it’s memetime. Until the internet dies.