When readers ask questions about my thriller novels, I love to discuss the themes and characters in them and the inspiration for my writing. But as generative artificial intelligence worms its way through the publishing industry, I’m bracing for a stomach-turning query: Did you actually write this?
The worry has been at the front of my mind since last week, when Hachette canceled the forthcoming U.S. publication of the horror novel “Shy Girl” after readers and journalists flagged prose that sounded like A.I. slop. (The author maintains that a freelance editor is to blame for any prose penned by a large language model.)
Though I’m against the use of generative A.I. in creative writing, not everyone feels the same way. What does seem clear, however, is that most readers want disclosure when A.I. has been used, and they are quick to note the telltale rhythms and patterns of popular large language models.
But as A.I. models continue to improve, I’m concerned that it will become difficult to distinguish between something written by a human versus a bot. As more A.I.-generated writing is put out in the world, more readers will question whether or not the text they are poring over was penned by a human. We’re barreling toward a rapid erosion of trust between authors and readers, and the publishing industry is unprepared to deal with the consequences.
Already, with a little fine-tuning, chatbots can be eerily good mimics of published writers, nailing their word choices and go-to grammatical patterns. James Frey, an author who’s no stranger to controversy and who has proudly admitted to using A.I. to write, has noted, “I have asked the A.I. to mimic my writing style so you, the reader, will not be able to tell what was written by me and what was generated by the A.I.”
Shortly after ChatGPT was publicly released, I entered the prompt “write a short story in the style of author Andrea Bartz.” The output was an uncanny facsimile of my prose — the actual scenes it generated made little sense, but the rhythm and sentences themselves mimicked some of the deliberate stylistic choices I make in my books.
A.I. detectors exist, but they’re far from perfect. OpenAI has called them unreliable. I don’t pretend to know how these checkers work under the hood. But if large language models were trained on my work (which was the case in at least one instance), then it’s easy to see how my own writing may come across to some as A.I.-generated.
In other words, I don’t write like A.I.; A.I. writes like me.
I pasted a few paragraphs of my own prose (a quick satire piece I’d shared on my Substack newsletter) into a free online detection tool. It deemed the passage “very likely A.I.-generated,” with 82 percent of the text exhibiting the hallmarks of A.I. This app appeared to be cruder and less reliable than other detectors I tried, perhaps because it was pushing a feature to “humanize” my passage with the click of a button. Lord help us all.
The relationship between artists and their patrons has always hinged on human connection. Though I have never met their authors, I cry when I read Sally Rooney’s “Intermezzo,” snicker my way through Kurt Vonnegut’s “Breakfast of Champions,” and feel my heart flip-flop when Mr. Darcy delivers his urgent line at the end of Jane Austen’s “Pride and Prejudice”: “If your feelings are still what they were last April, tell me so at once. My affections and wishes are unchanged; but one word from you will silence me on this subject forever.” Novelists write to wrestle with the human condition, to explore universal themes and alchemize their humanity into words that resonate with total strangers. If you remove the flesh-and-blood author from the equation — or plant seeds of doubt in readers’ minds about whether what they are consuming is authentically human — the writer-reader relationship falls apart.
In a worst-case scenario, mistrust could bring with it rampant accusations of A.I. use, including as a bad-faith tactic employed against enemies. By making bookworms paranoid about whether a poem or passage they love was written by a person or by an algorithm, it may also warp the act of reading itself. None of this bodes well for the future of publishing — an industry in decline, supported by fewer readers, paying authors unlivable wages.
I fear it could fall on us authors to prove our writing is our own. The advice I often hear from fellow authors is hopeful, if a bit hippy-dippy: Write something weird! Break all the rules! Pour your heart and soul into it because nobody can tell a story quite like you, baby! But in practical terms, there’s little writers can do beyond documenting the laborious process of penning a long-form piece. When it comes to academic writing and legal briefs, eagle-eyed reviewers can look for hallucinated citations, proof the work was A.I.-generated. For most writing, though, the primary indicator is alarmingly murky: Does it sound like ChatGPT to me?
Several made-by-human certifications exist for written works. But many of these programs operate on the honor system, so the badge lacks teeth. In theory, the author of “Shy Girl” could’ve slapped the Authors Guild’s “Human Authored” stamp on its cover (it doesn’t require detection tools). Book contracts typically state that authors’ work must be original. But again, enforcement options are effectively limited to fallible A.I. checkers that often have a high rate of false positives — and vibes.
The cancellation of “Shy Girl” raises many concerns about trust, authenticity and publishing’s readiness for a new, A.I.-assisted world. But one beautiful thing sprang from the fallout: Readers made it abundantly clear they want books by humans, not machines. This should hearten all the authors painstakingly getting words on the page one at a time — and frighten those secretly using generative A.I. as a shortcut to a finished draft.
Andrea Bartz is the author of “The Last Ferry Out,” “The Spare Room,” “We Were Never Here” and other novels. She also writes the Substack newsletter Get It Write.
The Times is committed to publishing a diversity of letters to the editor. We’d like to hear what you think about this or any of our articles. Here are some tips. And here’s our email: [email protected].
Follow the New York Times Opinion section on Facebook, Instagram, TikTok, Bluesky, WhatsApp and Threads.
The post A Horror Novel Got Canceled. What’s Coming Next Is a Bigger Nightmare. appeared first on New York Times.
