What's Happening?
Macmillan Publishers, a major U.S. book publisher, has discontinued its internal AI chatbot, NOA, after employees found it unreliable and largely unhelpful. The AI assistant, whose name stood for 'No Ordinary Assistant,' was launched in January 2024 with
significant corporate fanfare, including purple balloons and banners at the company's Lower Manhattan headquarters. Developed by the German software company Chaptr, NOA was intended to assist staff with tasks such as drafting emails, writing copy, conducting marketing research, and generating ideas. Despite being presented as a secure environment for proprietary material, employees were warned that NOA could 'hallucinate' and required human verification. By October 2025, Macmillan informed staff that NOA would be wound down. Subsequent attempts to integrate Google Gemini into publishing workflows also proved unsuccessful, with one editor describing its output for assessing literary agent submissions as 'useless.'
Why It's Important?
This development highlights a significant challenge in the U.S. publishing industry's adoption of artificial intelligence. While companies are eager to leverage AI for productivity gains, the practical implementation often faces resistance and dissatisfaction from employees who find the tools inconsistent or unhelpful. For Macmillan, a key player in the U.S. book market, the failure of NOA represents a costly misstep in its innovation strategy. It underscores the importance of user-centric design and thorough testing for AI tools, especially in creative fields where nuance and human judgment are paramount. The episode also reflects a broader tension between corporate enthusiasm for AI and the on-the-ground realities faced by workers, potentially influencing how other U.S. publishers approach AI integration and investment in the future.
What's Next?
Following the shutdown of NOA and the unsuccessful integration of Google Gemini, Macmillan Publishers will likely re-evaluate its AI strategy. The company may explore alternative AI solutions or focus on more targeted applications that address specific, well-defined needs within its publishing workflow. This experience could lead to increased skepticism regarding off-the-shelf AI solutions and a greater emphasis on custom-built tools or more rigorous pilot programs. Other U.S. publishing houses will likely observe Macmillan's experience, potentially leading them to proceed with more caution in their own AI adoption efforts. The broader industry may also see a shift towards AI tools that prioritize accuracy and reliability over broad functionality, especially in tasks involving sensitive or creative content.
Beyond the Headlines
The failure of Macmillan's NOA chatbot touches upon deeper implications concerning the role of AI in creative industries and the future of human labor. The initial celebration of NOA, contrasted with its eventual shutdown due to staff dissatisfaction, reveals a cultural clash between technological optimism and the practical realities of creative work. The label 'No Ordinary Assistant' itself, found offensive by some human assistants, highlights concerns about AI potentially devaluing human contributions. This incident could fuel ongoing debates about intellectual property protection in the age of AI, as well as the ethical considerations of using AI to generate content or evaluate creative submissions. It also raises questions about the long-term impact of AI on job roles within publishing, emphasizing the irreplaceable value of human creativity, critical thinking, and nuanced understanding in the literary world.
