In headline after headline this year, American book publishers have appeared to be fighting back against generative AI, from suing Google for copyright infringement to canceling books by authors suspected of using LLMs.

But behind the scenes, at least three of the Big Five US book publishers—HarperCollins, Simon & Schuster, and Hachette—have been quietly incorporating AI tools into their publishing processes without public disclosures or author consent.

According to interviews with more than two dozen workers, all of whom spoke under the condition of anonymity to avoid jeopardizing their employment, publishers are now using LLMs like Claude and ChatGPT to email literary agents, write publicity and back-cover copy, design cover art, and generate marketing materials, often at the direction of their executives.

At the same time, multiple editorial staffers say the widely publicized cases of authors accused of using LLMs to help write their books—like Hachette’s horror novel Shy Girl or Macmillan’s crime thriller Call Me, I’ll Hide the Body—are just the tip of the iceberg. “Every single imprint has a story of someone very esteemed who had a book canceled because of blatant AI use” without ever making the news, one editor says.

At HarperCollins senior leaders purchased Claude licenses from Anthropic months ago “but very quickly realized they didn’t know how to implement them,” according to a current employee. Dozens of junior- and senior-level staff were then “voluntold” to become “AI Champions” and were tasked with discovering use cases in monthly brainstorming sessions with staff across the company.

This spring, when some HarperCollins employees expressed concerns about the legal, ethical, and environmental impacts of AI at one of these brainstorming sessions, a person leading the meeting dismissed their qualms as “the cost of doing business,” according to someone who attended the meeting. Soon after, a statement was added to the new AI section of the HarperCollins online employee portal, noting that “while AI does have an environmental impact, it’s a small part of most individuals’ total digital carbon footprint.”

Some HarperCollins staffers were unnerved by the brainstorming sessions. “It was such a slap in the face,” says one employee. “We’ve been making valid business cases for years to be paid a healthy salary, and instead they spent that money on software that most of us don’t really want.” (The average entry-level salary for New York City publishing employees at the Big Five and Scholastic was $47,583 in 2023.) In addition to Claude, staffers say HarperCollins has also purchased licenses for ChatGPT and Jasper, an agentic platform that promises “to run end-to-end marketing workflows.”

At least one HarperCollins division has been “all in” on AI since gaining software licenses, according to one staffer, including AI-generated marketing videos. But according to workers at HarperCollins, Simon & Schuster, and Hachette, one of the most common use cases for LLMs so far is generating publicity copy.

Readers encounter publicity copy everywhere from Amazon and Goodreads to back covers, and it’s used to pitch books to booksellers, journalists, content creators, and critics. However, “using AI to generate publicity copy is short-sighted,” a Big Five publicist tells me. “Publicists have a duty to engage in good faith with readers and our colleagues in the media, who I’m sure have a low threshold for tolerating AI-generated pitches.”

On the other hand, Meg Reid, publisher of the independent Hub City Press, says LLM-written marketing and publicity unfortunately “feels pretty inevitable” for the Big Five. “That’s one place we’re probably going to see a bigger shift between corporate publishers looking for efficiencies and smaller publishers insisting that the humans involved are the whole point.”Meanwhile, earlier this year, literary agents began suspecting specific Big Five editors were using AI to write rejection letters.

“Among colleagues, we compared emails on different books from the same editors, and they all had the same classic signs of AI writing, like ‘not this, but that’ and the rule of three,” one agent says. Employees at Simon & Schuster, HarperCollins, and Hachette tell WIRED they believe the agents’ suspicions were correct.

A spokesperson for Hachette says the company “support[s] the use of AI tools for operational purposes, when they help our authors’ books reach more readers,” but that it “[does] not support creative uses of AI, including communicating with our authors and partners.”Wibke Grutjen, global chief marketing and communications officer at Simon & Schuster, says “employees across all functions have access to a limited number of enterprise-level AI tools,” but using them is not mandatory.

Grutjen said the company prohibits the use of any AI tool “that has not been vetted and approved” and that “the goal is always to free up more time for our employees to spend on deeper focus and creative work on behalf of and alongside our authors. HarperCollins did not respond to requests for comment.

Months ago, workers at Simon & Schuster attended a workshop with representatives from OpenAI. “The person leading the workshop genuinely had no idea what we did,” an attendee says, but one of the AI company’s recommendations for how to use ChatGPT in the book trade was, in fact, composing emails to literary agents. The practice has also grown among editors at HarperCollins and Hachette, employees say.

In May 2024, staffers say Simon & Schuster held a contest among employees to come up with the best use cases for AI, with a $10,000 prize for the winner and $5,000 each for two runners-up. Since then, the company has “rolled back some of the AI-pushing because employees kept complaining about it in town halls,” says one employee. According to several staffers, AI assets are also cropping up in cover art, with some foreign presses distributed by Simon & Schuster openly acknowledging AI use when sharing designs.

All of the agents I spoke with expressed fears that publishers might paste text from their authors’ manuscripts into the LLMs in order to generate things like rejection letters and publicity descriptions. “We’ve added AI clauses to our contracts saying ‘this material can’t be run through an LLM,’” says one agent, who did not want to be named to avoid professional repercussions. “Ostensibly, that was meant to protect authors from tech companies, but it also applies to the company buying and publishing the book. If they're not using a closed-loop model, they could be feeding an LLM with an author’s unpublished writing.”

Unlike the LLMs most people use online (Google Gemini, ChatGPT, et al.), a closed-loop model operates on an individual computer or network that’s “air-gapped” so it can’t share data with the AI company’s cloud servers. One source at HarperCollins says multiple editors have asked managers if they could paste a full manuscript into an open-loop LLM and were told not to by the company’s legal team.

Hachette employees are also bracing for the company’s AI policies to “evolve” under new chief information officer Carl Hixson, due to his previous role as chief technology officer at SRMG, a Saudi state-backed company that has publicly embraced agentic AI.

Some staffers across the Big Five say they’re reluctantly experimenting with LLMs only because executives gutted their teams through layoffs and subsequently tripled their workloads. “It’s skeleton crews everywhere you look,” says one employee.

“Large teams have been reduced to just five or six people, give or take, and we’re doing double or triple the number of books. No wonder people are taking shortcuts. I don’t judge them for that. I judge the company for shoving the shortcuts down our throats and saying, ‘Do this or we fall behind.’”

In just the past few weeks, word spread among Simon & Schuster workers that the company was conducting an experimental trial with Skan AI “at the apparent urging of KKR”—the investment firm that purchased Simon & Schuster in 2023. According to its own website, Skan AI “gives leaders the evidence base they need to reduce costs, prioritize automation, and meet board-level transformation mandates.”In response, Simon & Schuster employees began gathering signatures on an open letter to their CEO, Greg Greeley, expressing their “strong opposition” and “seeking a commitment to not bring Skan AI or other similar software into S&S workflows, either now or in the future.”

Beyond the software’s surveillance capabilities, many workers were also concerned with Skan AI’s “automation discovery” tools being used to shadow and replace them.

News of the open letter was published in New York magazine’s Book Gossip newsletter on October 6. On the same day, Greeley sent a memo to Simon & Schuster employees confirming that the company has been exploring Skan AI’s Blueprint product but that “no decision has been made to use Blueprint or any other productivity-monitoring or workflow-analysis tool,” and that he “had not seen the materials relating to this early-stage exploration” before concerns were raised to him the day before. Pangram, which, for better or worse, has become the go-to AI-detection platform, indicates that at least some of Greeley’s memo was written by AI.

Almost two years ago at the UBS Global Media and Communications Conference, HarperCollins CEO Brian Murray said AI presents “more opportunities than risks” for book publishers and that he saw AI playing a role in increasing the number of translated books and audiobooks published by HarperCollins in the future. Since then, multiple editors say they’ve seen LLM-translated manuscripts with AI-use disclosures at the annual Frankfurt and London book fairs, where US publishers buy international publishing and translation rights from foreign literary agencies.

In August, Michael Reynolds, executive publisher at Europa Editions, an independent press in New York best known for publishing Elena Ferrante’s Neapolitan Quartet of novels, told a room full of booksellers at the New Voices New Rooms conference in Baltimore he was excited about the possibilities for AI in translation and “seemed baffled by critiques of it,” according to attendees. Europa Editions is already controversial among literary circles for leaving the names of its human (for now) translators off the covers of its books.

“The C-suites at Big Five publishing companies have more in common with the C-suites at other major American companies than they do with the editors and publicists and translators who work for them,” says Dan Sinykin, author of Big Fiction: How Conglomeration Changed the Publishing Industry and American Literature. “They’re always looking for ways to maximize profit, and right now they’re trying to see what they can get away with before there’s a revolt.”

Most of the industry professionals I spoke with just want publishers to provide clear AI usage guidelines and disclosures, at the very least, to help mitigate the current chaos. “We need publishers to be clear about what is being used and when,” says Emma Dries, an agent at Triangle House Literary who previously worked in editorial at HarperCollins and Penguin Random House. A current HarperCollins staffer, meanwhile, says there has been no direction “about what to say to authors or agents about any of this, because I think they know it would start a shitstorm.”

“Short of these tools being banned, which would frankly be my preference,” Dries continues, “everybody should be communicating more.”