Big Five Book Publishers Quietly Deploy AI Tools as Staff Push Back

Major American book publishers are quietly weaving generative AI into daily operations even as some publicly criticize the technology. Interviews with more than two dozen workers show HarperCollins, Simon & Schuster, and Hachette have incorporated AI tools into publishing processes without public disclosure or author consent.

Quiet adoption at the Big Five

Staff say large language models such as Claude and ChatGPT are used to email literary agents, write publicity and back-cover copy, design covers, and generate marketing materials, often at executives' direction. At HarperCollins, senior leaders purchased Claude licenses months ago and then assigned dozens of staff as "AI Champions" to find use cases in monthly brainstorming sessions. When employees raised legal, ethical, and environmental concerns at one session, the leader dismissed their qualms as "the cost of doing business."

HarperCollins also bought ChatGPT and Jasper licenses, and at least one division produces AI-generated marketing videos. Staff across all three companies say generating publicity copy is among the most common use cases. Hub City Press publisher Meg Reid called LLM-written marketing and publicity "pretty inevitable" for the Big Five.

Staff revolt and surveillance fears

Many employees describe pressure to adopt AI after layoffs cut teams to skeleton crews and tripled workloads. "It's skeleton crews everywhere you look," one employee said. In May 2024, Simon & Schuster held a contest with a $10,000 prize for the best AI use case, but the company later "rolled back some of the AI-pushing" after complaints at town halls.

In recent weeks, Simon & Schuster workers learned the company was trialing Skan AI, a workflow-analysis tool, at the apparent urging of KKR, the investment firm that purchased the publisher in 2023. Skan AI says it helps leaders "reduce costs, prioritize automation, and meet board-level transformation mandates." Employees gathered signatures on an open letter to CEO Greg Greeley, who later confirmed the company explored Skan AI's Blueprint product but said no decision had been made.

Agents and authors demand transparency

Literary agents began suspecting some Big Five editors of using AI for rejection letters, citing repeated emails with classic AI patterns. Several agents added clauses to contracts stating manuscript material cannot be run through an LLM, fearing unpublished text could leak into open models. HarperCollins's legal team has told editors not to paste full manuscripts into open-loop LLMs.

A Hachette spokesperson said the company supports AI for operational purposes but not creative uses, including communicating with authors and partners. Simon & Schuster said enterprise AI tools are optional and limited to vetted software. HarperCollins did not respond to requests for comment.

Translation and the road ahead

HarperCollins CEO Brian Murray has said AI presents "more opportunities than risks" for publishers. Editors report seeing LLM-translated manuscripts with AI disclosures at the Frankfurt and London book fairs, and Europa Editions executive publisher Michael Reynolds has voiced enthusiasm for AI in translation.

Industry professionals are calling for clear AI usage guidelines and disclosures. "We need publishers to be clear about what is used and when," said Emma Dries, a literary agent at Triangle House Literary. Dan Sinykin, author of Big Fiction, said publishing executives are "trying to see what they can get away with before there's a revolt."

Sources: Wired Ai