Saturday, October 7, 2023

Lightlark, or, How I Finally Learned to Stop Worrying and Love the AI Takeover of Mass Market Fiction

I have been thinking a lot about the pending AI Fiction Apocalypse.

Artificial Intelligence has, by some accounts, already reached the point at which it can compose entire novels. At present they are barely readable. Very quickly, however, they will improve in quality, until they become better than all but the best human authors can compose. And then, at some point, they will be better than the best human work.

The publishing industry is, they say, looking forward to this, anticipating a world full of mass-market best-sellers that can be created with the press of a button, and a set of authors who don't have to be paid royalties. The fact that this future also renders the executives and publishers economically irrelevant has perhaps escaped them.


Alex Aster, I am told, is a TikTok personality who decided to write Lightlark on, well, a lark. Based on the jacket copy, she seems almost to have been designed in a lab for perfect YA marketability--she's an attractive white girl who graduated Summa Cum Laude from an Ivy League university and achieved celebrity in a community frequented by tween readers prior to composing her book.  If she didn't exist, the publishing industry would have to invent her.

The book itself was described by early readers in apocalyptic terms--as an absolutely incompetent clusterfuck. It's not quite that. The prose is florid bur readable. The story's world has some interesting elements to it. The characters are archetypes rather than fully formed individuals, but they're not completely unrecognizable as human beings. The story itself is an amalgamation of tropes and elements lifted from better books by more talented authors.  The twist at the end is an eye-roller but it doesn't cheat; it's consistent with the facts presented up to that point.

The book is not an enjoyable read by any stretch of the imagination--I had to force myself to finish it, like a toddler choking down his vegetables--but it's not unrecognizable as YA.  What it reads like, bluntly, is a YA novel as designed by a current-generation artificial intelligence. All of the ingredients of a mass-market best-seller are here, and they're assembled, as if by a template, in a way that sort of makes sense if you don't look at the whole thing with a human sensibility. 

A human will notice that the curses as described would result in the collapse of civilization within a couple of decades at most. A human will notice that the various ceremonies of the Centennial have nothing to do with solving the curses impacting the various kingdoms and everything to do with creating conflicts between their rulers. A human will notice that everybody is talking like a character in a 1930s movie serial. A human will notice that the visual imagery is assembled like a Mad Lib from a list of random metaphors. But at its current level, artificial intelligence might miss these things.

I am told that, in terms of sales, the book is a smash hit. Sequels are imminent. The film rights were sold before the ink on the initial print run was dry. 

The early, brutally negative reviews of the book have been buried an avalanche of five-star raves by young readers arriving from TikTok. I have always believed that the opinions of young readers matter more than those of critics where the quality of a YA novel is concerned. I have to assume that these kids have actually read and enjoyed the book, that they are not just review-bombing to support the Internet Celebrity Of The Moment.

As for the bestselling YA authors who chose to blurb this book...well, their behavior is less explicable, and less forgivable. I can't bring myself to believe that they actually read this and loved it. I don't really know how blurbing works. I am assuming that many blurbs are produced as a professional courtesy by pro authors who don't actually review the work they're blurbing.  So...ignore blurbs going forward, Steve. Lesson learned, I guess.

I do not _quite_ believe that Alex Aster is a fictional creation of the YA publishing industry, a name and an image they've attached to an AI model created to churn out genre fiction. I'm about 85-90% certain she's a flesh-and-blood human being. But I don't know how much it matters. This book probably wasn't written by a machine. But it might as well have been.

And, paradoxically, this has made me less afraid of the takeover of the fiction market by AI.

Because if Alex Aster _IS_ a living, breathing human, then the massive success of her book provides a model that YA publishers are going to emulate going forward. We're going to see the industry churning out more and more work by marketable social media celebrities. Ghostwriters will spackle over the more egregious cracks in the plot and slap a new coat of paint on the prose, and if the resulting product is never better than marginal in quality...well, so what? It won't matter. It will still sell.

There won't be any more JK Rowlings in this world, let alone any Madeleine L'Engles.  There won't be any reason to encourage original young authors who don't match the marketing profile. And as the market adjusts to these new realities, I suspect that young readers will re-calibrate their tastes as well. They literally will never know what they're missing.  Not unless they do a deep dive into the library stacks. And who has time for that?

So, in summary, Lightlark has helped me make my peace with AI-generated fiction. If we turn things over to the machines, our reading experience will improve over time. The human beings involved in producing YA appear to have things pointed in the opposite direction.

No comments:

Post a Comment