This is already happening for those in the know with public domain art. People like Jim Sterling lose their shit about asset flips, and rightly so, but they completely fail to notice stock textures and sound effects in even AAA games. I noticed some stock sound effects used in many games and films long ago, but now I'm a dev I'll be playing something like Resident Evil 4 HD and notice a stock wood texture. "Oh, that's just wood_04 from this free 2k texture pack. They didn't even try to hide it.".
There's even a
fascinating autistic fascinating subset of this called Texture Archeology where people go back the games from before high speed internet and try to source the textures.
Oh, yeah. That shit is amazing. And in fairness at least with audio some studios can do good work with stock effects.
Who'd have thought a scream that's been embedded into so many salty Souls players' forebrains came out of a silly little audio library?
I didn't know that was a thing. I thought publishing houses were inundated with stories at the best of times. It's why writing is such a hard career to break into. Why would they publish AI written books when there's countless people trying to get published but can't?
Note: This is just personal experience from my attempt at getting published many years ago (I ended up self-publishing, then hopped onto Amazon when they started their own self-publishing business), backed by talking to other authors since.
They haven't started doing it
yet, at least not that I know. There are people selling AI-written stories (usually short stories from what I gather) on Amazon and making a quick buck off uninformed idiots or people checking in for the meme, but it's a pretty small niche right now. AI is much, much better at short-form content like "news" articles.
Anyway, the problem for the publishing world isn't that 99% of the submissions they receive are shit. They're fine with shit. They publish shit all the time, because shit sells. The problem for them is going through all that shit and trying to find the shit that they think will sell. Because writers are humans, and even when a dozen aspiring writers are typing away trying to write the same flavor-of-the-month story that rides the coattails of the
last obscure-turned-viral award-winning story that propelled a bored housewife with unresolved fetishes into millionairedom,
they're probably all going to go about it in different ways.
So, it falls to the editors to go through all that shit and fish out the bits that they think will sell. Because the companies are generally not interested in good stories, they're interested in sales. If a good story sells, all the better. But it's not the pre-requisite there, at least not for the big guys in the game. They're there to make money, and the less they have to give the author, the better for their own bottomline.
Now, imagine that they then get access to an AI that they can direct to write whatever they want, in a way that's scientifically
designed to get mass market appeal. I don't think it's there yet (and even then, you'd need to hire someone skilled at wrangling the AI once it becomes available), but you can bet your bottom dollar the people at the top are salivating at the thought. They're unlikely to get the big blockbusters with this model, but getting 20 chances at success with AI-written stories for the same price as a single human-written story makes the beancounters very happy in their pants.
And for companies like Wizards of the Coast (to wheel this back to tabletop gaming), it just plain allows them to downsize their operations even further. Why keep a roster of writers that can only write 4 supplements a year, when you can keep only a few editors on hand and automate the production of 8 supplements for a fraction of the cost? Is it their prerogative as a private company? Sure. Doesn't mean I have to go along with it. Historically, companies maximizing profits has always resulted in worse results for their customers.
ETA: also, while I'm sure we're still a ways off wholly AI-written novels, there's an idea that I would be surprised if it weren't at least being tested as we speak. Namely, getting the AI to write the premise, characters and beats of a story, and paying a cheap author to hang "meat" onto that narrative skeleton (with the "author" getting credit if the resulting product becomes successful). Or using AI to embellish/pad out dry prose. At the end of the day, it's a tool for reducing workload (and therefore cost). It's possibilities are endless for a company in search of profits.
