Three of the five largest publishers use Claude and ChatGPT without informing their authors
At least three of the five largest American publishers use ChatGPT and Claude for everything from emails to literary agents to jacket copy and marketing, according to WIRED's investigation. Employees are now starting to push back – and criticism is also targeting surveillance tools linked to a publisher's owner.
According to an investigation published by WIRED on October 9, 2026, at least three of the "Big Five" in American publishing – HarperCollins, Simon & Schuster and Hachette – have adopted AI tools in their publishing processes without disclosure or authors' consent (WIRED). The basis is interviews with more than two dozen employees, all of whom spoke on condition of anonymity to avoid jeopardizing their jobs.
How the AI Is Used in Practice
The anonymized employees report that the publishers use language models such as Claude and ChatGPT for a range of tasks: writing emails to literary agents, producing review and back-cover copy, designing cover art and generating marketing material – often on direct instruction from management. None of the publishers has individually confirmed or denied these specific uses; what the companies have offered are general policy statements.
The idea that usage is voluntary is thus contradicted by a different picture internally.
HarperCollins: Licenses Without a Plan
The HarperCollins case illustrates how the usage has developed. Senior leaders purchased Claude licenses from Anthropic several months ago, but quickly realized they did not know how to put them to use, according to a current employee WIRED spoke with. Dozens of employees at all levels were then – in the source's word, "voluntold," a pun on "volunteer" and "told" – designated "AI Champions," tasked with finding use cases through monthly brainstorming sessions across the company.
There are also internal limits – but they have been kept to themselves. According to a source at HarperCollins, editors who asked whether they could paste an entire manuscript into a consumer LLM tool were told no by the legal department. The rule has never been made known to the people who wrote the books.
The Rejection Letters: Suspicion, Not Proof
Earlier this year, literary agents began suspecting particular editors at the big publishers of using AI to write rejection letters. "Among colleagues, we compared emails about different books from the same editors, and they all had the same classic hallmarks of AI writing, like 'not this, but that' and the rule of three," one agent told WIRED.
The distinction here must be clear: these are stylistically based suspicions, not documentation such as email metadata or admissions. But the employees themselves believe the suspicion may be right – employees at Simon & Schuster, HarperCollins and Hachette tell WIRED they believe the agents are correct.
The Pushback Goes Public
Friction inside the organizations broke into the open in October. Employees at Simon & Schuster circulated an open letter to chief executive Greg Greeley, a former Amazon executive who took over in March, opposing a pilot program with Skan AI – a workflow-monitoring and automation tool that employees linked to KKR, the investment firm that bought the publisher in 2023 (Martin Cid Magazine).
On October 6, Greeley responded to employees that the company had explored the product, but that no decision had been made to adopt it or a similar surveillance tool.
It is worth keeping in mind how extensive the pushback actually is: WIRED's headline ("Staff Are Revolting") suggests a broad movement, but the documented starting point is primarily the single open letter at Simon & Schuster. And Martin Cid Magazine's article, which expands on the letter and Greeley's response, draws on the same WIRED material – it is not independent confirmation.
The Publishers' Official Lines
The companies respond with qualifications rather than denials. A Hachette spokesperson says the company "supports the use of AI tools for operational purposes, when they help get authors' books out to more readers," but "does not support creative uses of AI, including communication with our authors and partners."
Wibke Grutjen, global head of marketing and communications at Simon & Schuster, says that "employees across all functions have access to a limited number of enterprise-level AI tools," but that use is not mandatory.
The Road Ahead
Usage thus appears to be moving deeper into the value chain. HarperCollins chief executive Brian Murray has said AI could help the company publish more translated books and audiobooks. Editors tell WIRED they have seen machine-translated manuscripts – labeled as such – offered for sale at the Frankfurt and London book fairs, where rights to foreign titles are traded.
Several questions remain open. No primary documents have been published; the entire picture rests on anonymous interviews and the companies' own statements. It is unclear how widespread internal resistance is beyond the open letter, and none of the three publishers has taken a position on the specific uses the sources describe. For authors, the most fundamental issue remains: neither the purchase of Claude licenses, the "AI Champions" program nor the legal department's manuscript rule has been communicated to them.