Major US publishers quietly adopt AI tools despite public anti-AI stance

Internal adoption remains hidden

HarperCollins, Simon & Schuster, and Hachette have incorporated AI tools into their workflows without telling the public or getting author consent. Interviews with more than two dozen anonymous workers reveal that executives direct staff to use large language models for emails and marketing materials. These companies purchased licenses from providers like Anthropic but struggled initially with implementation before assigning tasks to junior and senior employees.

Staffers describe secret usage

Employees report using Claude and ChatGPT to write publicity copy, design cover art, and generate marketing content across multiple imprints. One editor stated that many esteemed authors had books canceled for blatant AI use without the stories ever making national news. This internal practice contrasts sharply with headlines showing publishers suing Google or canceling books by suspected users of generative models.

Workers speak anonymously

All interviewed staff members agreed to anonymity to avoid jeopardizing their current employment situations with these major publishing houses. The reports indicate that the publicized cases involving specific novels are only a small fraction of the total number of cancellations occurring internally.