Macmillan Staff Say the Publisher Threw a Party for a Workplace AI Assistant That Was Later Killed
wired ·
Macmillan staff reported that the publisher launched an AI assistant named NOA with a celebratory party in January 2024, only to discontinue it later due to its unreliability and lack of utility. Employees described the tool as offensive and unhelpful, noting that despite training warnings about hallucinations, they were encouraged to use it for drafting emails and generating performance review goals. By October 2025, Macmillan sunsetted NOA and began experimenting with Google Gemini, though internal feedback suggested these new tools were also ineffective for creative tasks. This incident reflects a broader tension within the Big Five publishers, where executives are pushing for AI adoption while simultaneously suing AI companies for copyright infringement. Bestselling authors have expressed frustration over this hypocrisy, citing issues such as AI-generated marketing copy containing hallucinated quotes and editors failing to read manuscripts. Many writers feel that the industry's reliance on AI undermines trust and devalues human creativity, leading some to consider self-publishing instead. Despite corporate statements emphasizing thoughtful integration and employee education, staff members often avoid using these tools because they hinder rather than help their work. The growing disconnect between publisher strategies and author expectations is causing significant strain in professional relationships across the publishing sector.
wired