HarperCollins, Hachette, S&S staff push back on AI rollout

Workers at three of the five biggest publishers tell Wired that chatbots are now drafting publicity copy, cover art, marketing videos, and emails to literary agents, in most cases without public disclosure or the consent of the authors involved.

Adam Morgan's reporting draws on more than two dozen employees at HarperCollins, Hachette, and Simon & Schuster. Most described the usage in terms like "gross," "unfortunate," or "short-sighted," while acknowledging that colleagues were reluctantly experimenting after executives slashed their teams and increased their workloads. At HarperCollins, senior leaders bought Claude licenses from Anthropic, and dozens of staff were drafted as "AI Champions" and asked to find uses in monthly brainstorming sessions. At Simon & Schuster, representatives from OpenAI ran a workshop for staff and suggested ChatGPT for composing emails to literary agents, and the company ran an internal employee contest with a ten-thousand-dollar top prize before reportedly easing its push after complaints.

Three of the AI practitioners we track in our Who's Who directory posted the piece shortly after it ran, unusual pickup for a trade-press story.

The sharper edge in the reporting is a monitoring tool. Simon & Schuster explored Skan AI's Blueprint, software that analyzes workflows and can identify automation opportunities, and employees are collecting signatures against its deployment. Hachette, by contrast, told Wired it supports operational AI but not creative work such as communications with authors and partners, a position the reporting itself shows is porous in practice.