I don’t like making predictions about AI and publishing. First, so much remains undiscussed and unknown: few publishers or professionals are entirely forthcoming about AI use. Second, the technology has progressed rapidly in the last year, now possessing capabilities I would’ve never expected. How much more will it progress in the near term? I find it impossible to say.
I’m fielding 1–2 media calls per week to discuss AI and the publishing industry, everyone seeking answers, yet thequestions will outpace the answers for years to come. However, the more I discuss AI detection and allegations of AI use against authors, the more I think this particular problem will have a short shelf life. Here’s why.
I believe the works now caught up in AI allegations were written, submitted, edited, and published before the industry came to fully understand how many authors use AI without disclosure. Until recently, few publishers were looking for such use or expecting it, especially on the fiction side. (University and academic publishers are a different story.)
If I were a creative writer using AI
If I were a creative writer using AI to assist me with my writing prior to 2026, and I decided not to disclose my use, I might have felt confident in my ability to pass off the writing as my own. Prior to 2026, writers didn’t ask me questions about AI detection; instead, writers asked me to how to protect their work from AI training or prevent their publisher from using AI in the creative process. So I would argue that prior to Pangram coming to the fore—specifically in the Shy Girl case—fiction authors in particular have been assuming there’s no way to prove they did or didn’t use AI. And that’s probably the best position to take if you’re accused because very few people are going to stand up and defend AI detection software in the current environment. Anyone who finds Pangram useful has to, in the same breath, also talk about how it’s not always right. But it’s rightoften enough to be useful to publishers and agents, and that’s why these scandals will end.
I didn’t hear consumer publishers or agents say they were using AI detection such as Pangram until earlier this year. Few writers, I would argue, had heard about Pangram until the last nine months. (If you were a reader of my paid newsletter, you would’ve seen me report on Pangram for the first time in 2024 but not again until 2026.)
Anyone paying attention must realize that, going forward, agents and publishers can’t take writers’ word for it that their work is their own. Sure, there’s a lot of nice talk about publishing being built on trust, but let’s be honest: That’s nonsense when you’re a publisher working at scale or with authors with whom you don’t yet have a trusting relationship. So work is going to be scrutinized and, yes, scored by Pangram or whatever tools become available to help publishers and agents assess the qualities of the work before them. Like it or hate it, they’re doing it for everyone’s protection. It benefits nobody—not author, publisher, editor, or agent—for an AI accusation to be made by the media, by a journalist, by an anonymous social media account, etc. Any smart publisher is going to get ahead of those accusations and be aware of a discoverable score prior to publication. Any agent submitting to a publisher is also going to get ahead of those accusations. That’s why I believe these public, surprise accusations will eventually become nonexistent because publishers and agents are now paying attention and they won’t move forward with a project that might prove troublesome in its provenance.
2026 is a unique year then
2026 is a unique year then. Many of the books now under scrutiny probably went through agents and editors in manuscript form in 2024–2025, before the industry was really thinking about AI writing or had any methods of identifying AI-assisted writing. By the time those books came to market, detection tools had improved and were being widely adopted. So the scandals of 2026 come from the gap between those two moments.
Writers will adapt. They now know they can’t use AI to write large swaths of a work without doing something to make their Pangram score acceptable. Again, like it or hate it, it’s now the path we’re on, until such a time that AI advances beyond the reaches of Pangram, when the dividing line between AI writing and human writing has become so fuzzy, it’s a useless distinction. Some believe there will be a technical means to establish human provenance for a work, like some highly sophisticated method of keystroke capture or file saving that’s powered by the blockchain. (Don’t ask me what that all means, I’m just sharing the services that get pitched to me all the time.) But no one in publishing has the patience (or money or time) for that.
I’ve seen many people suggest that authors accused of AI allegations should sue or will sue—or that lawsuits will somehow bring AI detection to an end. I do expect court cases that seek to prove how much of a work is human versus AI, but I think that’s most likely to happen in relation to the question of copyright. Right now, a work has to be human authored to receive full copyright protection and the US Copyright Office is starting to ask questions about how authors used or didn’t use AI. There have already been lawsuits brought by authors/creators seeking copyright for works they created with the aid of a machine. That’s the issue I expect we’ll be fighting over, not AI detection.
Jane Friedman has spent her entire career working
Jane Friedman has spent her entire career working in the publishing industry, with a focus on business reporting and author education. Established in 2015, her newsletter The Bottom Line provides nuanced market intelligence to thousands of authors and industry professionals; in 2023, she was named Publishing Commentator of the Year by Digital Book World.
Jane’s expertise regularly features in major media outlets such as The New York Times, The Atlantic, NPR, The Today Show, Wired, The Guardian, Fox News, and BBC. Her book, The Business of Being a Writer, Second Edition (The University of Chicago Press), is used as a classroom text by many writing and publishing degree programs. She reaches thousands through speaking engagements and workshops at diverse venues worldwide, including NYU’s Advanced Publishing Institute, Frankfurt Book Fair, and numerous MFA programs.
