Book publishers’ organizational use of AI rose in a 2026 North American industry survey, even as individual respondents reported using it less. That divergence points to growing workplace caution and changing practices—not proof of a coordinated staff revolt, or that publisher rollouts caused the decline in individual use.
What the 2026 survey says
The survey, conducted across the United States and Canada, drew 771 industry professionals, according to Publishers Weekly’s account of the 2026 results. It is a survey, not a census of publishing companies or workers.
As an Amazon Associate I earn from qualifying purchases.
| Measure | 2025 | 2026 |
|---|---|---|
| Publishers reporting organizational AI use | 48% | 63% |
| Individuals reporting AI use | 46% | 38% |
| Organizations reporting an official AI policy or guidelines | 31% | 49% |
The organizational and individual figures are reported by Publishers Weekly’s 2026 coverage; the policy figures come from the Book Industry Study Group (BISG). The year-over-year pattern is clear: more organizations said they use AI, while fewer individuals said they do. These are separate survey measures, and the results do not show why they moved in opposite directions.
Recommended Free Tools
Where publishers and workers report using AI
Reported use is concentrated in work that supports publishing operations, rather than evidence of a broad move to AI-written books. BookNet Canada’s 2026 breakdown reports the following shares for three common task categories:
#1 Best Overall
| Task category | Organizations | Individuals |
|---|---|---|
| Administrative or operational tasks | 56% | 46% |
| Marketing | 54% | 44% |
| Data analysis or reporting | 41% | 36% |
The survey material also describes research and metadata-related activity, but these findings do not establish that publishers generally use AI to write books. They show reported use across task categories, not how often a tool is used, which systems are involved, or whether a particular task is handled entirely by AI.
More policies do not mean uniform adoption
Nearly half of organizations reported an official AI policy or guidelines in 2026, up from 31% in 2025, according to BISG’s survey summary. BISG describes the most common organizational approach as allowing AI in controlled workflows or for specific tasks. That is different from unrestricted use across a company.
Policies and actual use are not interchangeable measures: a policy may define permitted tools or tasks without indicating how widely employees use them. Adoption also varies by sector and organization size, as BookNet Canada’s industry research page notes. The survey therefore describes a mixed landscape, not a single publishing-wide rulebook.
Does “staff are revolting” accurately describe the findings?
The headline’s wording is stronger than what the survey establishes. The 2026 results show declining individual use alongside increasing organizational use; they do not document coordinated protests, strikes, or another industry-wide revolt. Nor can a year-over-year comparison prove that employer adoption caused individual use to fall.
Rank #3
There is evidence of concern. In its account of the 2025 BISG survey, Publishers Weekly reported that nearly half of professionals used AI for work and 98% reported significant concerns about implementation. Those figures belong to the 2025 survey and should not be read as 2026 results. Together with the later decline in reported individual use, they support a picture of caution and debate—not a proven, coordinated revolt.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Why the industry debate is bigger than workplace tools
Workplace experimentation sits alongside two distinct disputes: whether AI-detection tools can reliably assess submissions, and whether publishers’ copyrighted books have been used to train AI models. Publishers Weekly’s September 2026 feature discusses these issues alongside internal experimentation. They help explain why AI policy is contentious, but they are not measurements of employee opposition in the survey.
Rank #4
The feature also cites Bertelsmann’s January 2025 announcement that it deployed ChatGPT Enterprise for employees and used AI in marketing and demand forecasting. That is one named-company example, not evidence of a standard approach across publishers.
Quick wins for a faster PC:
Scan for outdated or missing drivers - takes under a minuteDriver Scan →Repair Windows errors before they cause bigger problemsFix Now →Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
