Writing Books in the Age of AI
Author Laura Brooke Robson breaks down the turbulence of writing books in the age of AI and proposes the answer may be to write weirder.
Right. Obviously! You know this to be true because I just made a little staircase with my sentence, like I am 19 in a post-modernist poetry seminar. You know it’s not AI writing because I don’t sound weirdly semi-formal, like the human embodiment of jeans and a blazer. You know it’s not AI because this is Writer’s Digest.
Then again.
I’ve started reading like a detective. I do not want to do this. I want to sit on my park bench with my little thermos of coffee and fall so completely into the book I’m reading that I forget everything else. But lately, I’ve had a hum in the back of my head. This hum, it’s constantly debating whether the thing I’m reading was actually written by AI. Emails, books, captions, articles. Sometimes, I don’t really care. Sometimes, I really do. But no matter how trivial, I can’t shake the hum, because I want to understand how good AI is getting. I want to know if I’m a better writer than the machines.
I live in New York, so most of my friends work in either tech or the arts. My tech friends were quick to fall in love with AI; my arts friends are steadfast in their resentment. My take on this is that AI is so unambiguously good at coding that it’s extremely difficult to resist it if you work with code. So all the tech folk say, “Yes, wow, we do need to be afraid! AI is amazing!” And all my arts friends, all my book friends, are reading the literary output of ChatGPT and saying, “This? You think this is amazing?” Because AI is better at coding than writing books.
When a tech person tells me AI is great at writing, it makes me feel defensive and possessive. It makes me annoyed that they can’t, or don’t want to, see the nuance in a piece of fiction that makes it really great, truly amazing. And I think that’s what so many of us are afraid of when we talk about AI. Not that we will be fooled. We’re all each too clever for that, aren’t we? No: We are afraid everyone else will read their AI books, and they won’t notice they were written by AI. Or worse: They will know, but they won’t care.
Every few months, I check in on Claude and ChatGPT to see how their fiction skills are coming along. They’re getting better. I really don’t love saying that. They totally are, though. A year ago, a piece of dialogue might have looked like this:
“I don’t want to!” Ralph shouted loudly, stomping his foot.
“You must!” Ralph’s mother opined, wiping the sweat from her brow.
“No!” Ralph extolled, crossing his arms.
*****
“Aha,” I relished, nodding smugly. But now, a piece of dialogue might look like this:
“You can’t make me go, Mom.” Ralph looked away, his lips pressed together. “I’m not a child.”
His mother took a step forward. “I’m not saying you’re a child. I just thought, since this matters so much to your father–”
Ralph let out a sharp laugh. “Don’t talk to me about my father.”
*****
“Hmm,” I say, feeling less smug. Is it wonderful? No. It still skews toward the choppy (I don’t like my dialogue broken up with this much narration) and verbose (why “let out a sharp laugh” when you could say “laughed”?). But it’s a lot better than it was. If version 1 is the way I wrote when I was 10, version 2 is the way I wrote when I was 20. Just doing some quick math here, but that may indicate AI writing in one year’s time will be the quality I could write by age 30. As I am currently 29, this bodes very poorly indeed.
I love to write. I want to keep writing as long as readers will let me. When I look at AI getting better at this whole prose thing, I feel a pressure in my chest, like someone has grabbed my heart and squeezed. I want to be a better writer than AI. I want my prose to be more interesting, more lovely, more full of feeling. I hope it can be. But I also have developed a twinge of fatalism. The truth is, I’m not convinced I can be better. But I am fairly sure I can be weirder.
My book, Love Is an Algorithm, is about AI. Specifically, it’s about the love story between the co-founder of an AI dating app and a musician trying to make a living as an artist in the era of LLMs. So you can tell all this stuff was on my mind as I was writing it. I kept asking myself, “Could AI write this?”
As a result, I made some weird formal choices. This, for me, was exceedingly fun. The chapters all have slightly different formats. Some jump back and forth through time; others are interspersed with Slack messages and scraps of terrible lyrics. At the time of writing, I’d started reading so much AI-generated content (mostly in emails and on social media) that I’d found myself getting bored and distracted more easily.
Because AI content is boring, isn’t it? The nonspecificity of it, the frictionless vagueness, gives you nothing to hold onto. I hate being bored, and so that became the challenge and joy of writing this book: Don’t get bored. If you feel like you’re writing yourself into a rut, do something weird. Veer in a new structural direction. Invent a word, make a joke, surprise yourself. And the book really was a joy to write. And through the experience of thinking so much about weirdness, I have noticed that the weird is all around. In fact, I think weird books are having a bit of a moment. I think we are all hungry to experience art that feels surprising—and human.
One of my favorite books so far this year was Lost Lambs, by Madeline Cash, which I chose as my Book of the Month book (shoutout Book of the Month). In the first few pages, you learn that there is a gnat infestation in the church. Cash made the really delightful decision to bring this infestation to life by spelling “naturally” as “gnaturally.” I mean. Love! This is great stuff. ChatGPT is absolutely not dropping the word “gnaturally.”
Another favorite: Heart the Lover, by Lily King. This book covers a lot of time, and a less interesting, less wonderful writer than King might have made it firmly chronological, or might have leapt back and forth between the past and present threads with each chapter. But at no point in my reading of this book was I sure exactly when the story would take me, and I loved every moment. I got to know exactly the information I needed to know, when I needed to know it. I have yet to see an AI-generated outline that can be so deft—and build a structure so deferential to the reader’s experience.
One more, if you’ll allow me: I loved Margot’s Got Money Troubles, by Rufi Thorpe. It’s so funny. Its humanness is in how completely hilarious and genuine it is. And comedy and surprise are so tightly linked: Nothing will make me laugh like something I did not expect to hear. And AI has yet to write prose that surprises me.
It’s not just books. I like hearing podcasters mispronounce words, laugh, and correct themselves. I like when someone fires of a zany, too casual email in a work context (“best regardz”). I even kind of like seeing typos in the wild lately. In a moment when we’re all being asked to question our realities all the time, I find it immensely soothing to know I am in the presence of humans.
I keep thinking back to this writing competition I judged a few years ago. Generative AI was younger but already accessible. Sometimes, I would start reading a piece, and it would be clean and approachable, so I would think, “Seems pretty good.” That was the first check my brain was doing on the writing: Can I understand this?
And then I would read a little bit more, and my eyes would glaze over. I would chide myself and go back to the start. Focus up, kiddo! I would reread, and I would try to sink into the story, and I would feel a slow fear creep up the base of my neck. This, I thought, was probably AI.
It’s a funny thing, to find yourself unwittingly reading AI prose. It’s a fool’s errand to pretend we can squish AI out of existence, or that no good can come of it. But it is hard not to feel betrayed when it has been foisted on you.
So I would keep reading these entries, now wary, afraid someone was trying to pull the electric sheep’s wool over my eyes. But then I’d see something: Something weird.
I’d see something weird, like a typo, or an invented word, or a shift in form and structure. I’d see a joke that caught me off guard. I’d catch a glimpse of the messy human behind the curtain, and I’d think: Ah. And only then could I relax and really, truly read.
Check out Laura Brooke Robson's Love Is an Algorithm here:
(WD uses affiliate links)









