So it looks like there's going to be a writers' strike in the United States. Among the demands brought forward by the Writer's Guild of America is the regulation of "generative AI" in screenwriting: the use of large language models like GPT, which produce text by calculating where certain words in the English language are statistically most likely to appear next to each other.

No matter what your job is, there's an AI booster out there who thinks GPT can do some part of it better than you can. Those guys are frequently wrong; for example, here's a post by Bret Devereaux examining in-depth the idea that ChatGPT can write your college essays for you. Short answer: it can produce an assemblage of text that looks like an essay, but submitting that text as your essay will not result in a good grade, because a truly successful essay requires a cognitive depth that is completely beyond large language models.

Where ChatGPT and other generative text models like it actually excel is in producing text that meets formulaic requirements in a confident, passable, vaguely novel and blandly inoffensive format.

It is, from a certain point of view, the perfect Hollywood screenwriter.

Book cover for Save the Cat!

Have you ever wondered why so many movies feel the same, lately?

In 2005, Blake Snyder published Save the Cat! The Last Book on Screenwriting You'll Ever Need. It took the three-act structure popularized by Syd Field and broke it down even further into a series of common story beats, even going so far as to lay out specific page numbers on which each beat should occur.

Whether Save the Cat! created a formula or merely formalized it is debatable. But for many Hollywood movies, there's some version of the Save the Cat! beat sheet in play. You can even buy Save the Cat!-branded story structure software.

If you're a screenwriter working on one of these movies, your role is not one that requires passion, inspiration, or any kind of sincerity. Your role is to produce content that fits the requirements of each story beat in a form just barely novel enough to avoid intellectual property infringement.

You are a human doing the work of an algorithm.

Google search for 'dean winchester death no no not that one.' Below it is a suggested search: 'dean winchester daddy issues.' Below that is a request to report any inappropriate predictions.
No, I'd say that prediction is entirely appropriate.

The last time the WGA went on strike was in 2007, lasting into 2008. Multiple shows abruptly finished ongoing seasons—perhaps the most notorious example being Supernatural, which decided to end season 3 early by (temporarily) killing off one of the main characters and sending them to hell. The strike also resulted in a deluge of unscripted "reality" shows and imported foreign TV series to fill out the broadcast schedule.

This time around, I think America's big production companies—Disney chief among them—will see GPT as a strike-breaking tool. Rather than another flood of reality TV, it's entirely likely we'll start seeing movies and television episodes "written" by large language models while producers refuse to negotiate with the WGA and human screenwriters remain on strike.

I don't think this is a good thing. I just think this is an entirely probable thing. Large media companies aren't interested in concepts like "art" or "basic standards of living," they're interested in making money. And if they can't persuade you to buy a separate Netflix subscription for everyone in your family, they'll instead cut costs wherever they can. If that means putting human screenwriters out of business, they'll do so happily.

Once upon a time, when generative text models were slightly worse at their job, this might have resulted in some delightfully strange movies. Nowadays, what we'll probably get instead is the audiovisual equivalent of nutrient slurry.

Which, admittedly, is where things were already headed.

style credit