Is AI the Beginning or the End?
I’ve been deluged recently with information about various AI (artificial intelligence) programs and how they can help writers. Or replace writers. Or anything in-between. The venerable Authors Guild devoted most of its latest member bulletin to the subject and to the analysis of various programs. And this week, when I sent my publisher the beta-draft version of my new mystery, one of the things I received in return was an AI analysis of the novel—which I am still digesting.
In 2016 a Japanese novel came close to winning a literary award. Nothing new there—except that it was written completely by a computer. A 2020 article in the Guardian was also completely AI-generated.
I’m not worried—yet. AI can’t do what I can do—yet. Computers can’t just create stories in a void; instead, they’re given a massive number of exemplars—works that are similar to what the computer is tasked to create—which is then broken down into data. AI reads the data using statistical algorithms to recognize patterns and relationships.
The Guardian article instances something called GPT-3, a text-generating AI tool (it’s actually a neural net trained to write by consuming massive amounts of text from digitized books across the web) that can create fairly sophisticated material. In 2021, Stephen Marche wrote in The New Yorker that Sudowrite (an AI application based on GPT- 3) was able to continue The Metamorphisis where Kafka left off.
That’s a little scary… but really only a little. As noted by AG president Mary Rasenberger:
These AI-created “new” works, as such, are not really new or truly creative. Rather, the AI system ingests existing works that have been broken down into data points, and then rehashes them — albeit in a very complex way. As Dr. Alison Gopnik, a professor of psychology and part of the AI research group at the University of California, Berkeley, puts it: “We might call it ‘artificial intelligence,’ but a better name might be ‘extracting statistical patterns from large data sets.’” While some might argue that the human brain creates in that same way, and certainly there are examples of books, songs, and so on that merely rehash what already exists, it is clear that humans bring something more to what we would consider truly creative works — their experience, emotion, imagination, and intuition. Human art evolves because it strives to reflect the stories of our time and place. Computer art will always stand still without new human art to train on.
A writer/editor/techie I follow online is called Renaissance Rachel, and she has put together an exhaustive list of various current (2022) AI programs for writers. It’s worth checking out, if only to see how far we’ve come from the days of ELIZA in MIT’s lab—and worth noting, as well, that the aim of the ELIZA enterprise was to demonstrate the superficiality of communication between humans and machines. We’ve come a long way since then, though it has to be noted that while the program was created in the 1960s, it was still being used by—in my opinion—way too many psychologists for prescreening patients as late as the 1980s!
The analysis my publisher sent me was full of data, some of which feels irrelevant, much of which feels valuable—though I’m not yet sure in what way. Back in 2005, screenwriter Blake Snyder wrote a bestselling nonfiction book on the craft of screenwriting titled Save The Cat. Since that time there have been a number of sequels, including for our purposes Save The Cat Writes a Novel.
The title refers to a particular plot device in which the protagonist does something admirable (like saving a stranded cat) toward the start of the story in order to establish the protagonist as a likable person and get the audience on their side.
I read Save The Cat Writes a Novel. I even sat in on a webinar presented by the author. And I ended up shaking my head and putting it away, because the Save the Cat method is so… formulaic. Save the Cat breaks your story down into fifteen key moments, even explaining at what percentage of your novel that beat should occur. The beat sheets are divided into three acts, which are then further divided (are you still with me?) into the corresponding beats. You can download versions of the beat sheet on various websites and it plots popular movies against the beats to show the method’s application in successful products. You can also try the beat sheet calculator, where you enter the page count of your novel; it will outline on which page you should be hitting each of your beats.
Once upon a time, I worked as a technical writer, and I can just feel technical writing permeating this novel framework. I don’t want to dismiss it—the numbers don’t lie, and obviously many successful novels and screenplays have been created and enjoyed through the use of the Save the Cat methodology.
But my stories grow organically, even the mysteries—maybe especially the mysteries—through creating characters who feel real and who interact with each other in the ways people interact in the world. You would never have plotted your life the way it’s unfolded, right? It’s my sense that staying close to real life will keep stories real and relevant, too.
Which brings me back to AI and its Save-the-Cat version of storytelling. I actually rather enjoyed looking at the analysis the program provided of my novel, possibly in part because most of the categories weren’t issues I had thought much about. That may be because I’ve developed over the years an inner sense of pacing, and dialogue, and characterizations, so I no longer think about them. It may also be that I’m a moron. It could go either way.
But it was still fun seeing what the machine thought of my story. We’ll see where it all leads as my beta readers get back to me and I start on my (hopefully) final revisions!