Publicly, American book publishers project an image of aggressive resistance against generative AI, launching lawsuits against technology providers and penalizing authors suspected of utilizing large language models. Behind closed doors, however, operational reality dictates a starkly different strategy. 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.
Publishers are now deploying models like Claude and ChatGPT to email literary agents, draft publicity and back-cover copy, design cover art, and generate marketing materials. The financial motivation for this internal shift is direct. In an industry where the average entry-level salary for New York City publishing employees at the Big Five and Scholastic sat at $47,583 in 2023, leadership teams view software automation as a necessary lever to compress operational overhead, purchasing enterprise licenses for platforms including Claude from Anthropic, ChatGPT, and Jasper.
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.
Execution has consistently lagged behind executive ambition. At HarperCollins, senior leaders purchased Claude licenses months ago only to realize they lacked internal deployment frameworks, subsequently forcing junior and senior staff into brainstorming sessions to discover practical use cases. When employees raised legal, ethical, and environmental objections during these sessions, management dismissed their concerns as a baseline cost of doing business, treating institutional friction as a mere nuisance to be bypassed rather than a genuine operational risk.
Operational Friction and Internal Resistance
The internal push toward automation has created severe friction between corporate management and editorial staff. Employees who fought for sustainable compensation find themselves pressured to train the very systems targeted to replace routine publishing tasks, such as generating promotional pitches and publicity copy. Literary agents began suspecting specific Big Five editors of using AI to write rejection letters after observing recurring stylistic markers across correspondence from identical editors, a suspicion corroborated by internal workers at Simon & Schuster, HarperCollins, and Hachette.
Corporate communication strategies attempt to draw a sharp line between administrative efficiency and creative execution. As a spokesperson for Hachette stated, the company supports the use of AI tools for operational purposes when they help authors' books reach more readers, while explicitly prohibiting creative uses including communication with authors and partners. Similarly, as Wibke Grutjen, global chief marketing and communications officer at Simon & Schuster, noted, employees across all functions have access to a limited number of vetted enterprise-level AI tools that remain non-mandatory.
Deploying language models for client-facing tasks like publicity copy and agent correspondence introduces clear commercial risks for marginal labor savings. Publishers attempting to cut administrative overhead through automated workflows are systematically eroding the relational trust required to maintain a competitive market position against independent alternatives where human curation remains the core product value. When the primary differentiator of an industry is taste, outsourcing the first impression to an algorithm looks less like digital transformation and more like a slow-motion institutional liquidation.