Since generative artificial intelligence (AI) models hit the big time in 2023, many involved in scholarly communications have pushed for rules and policies around how authors and peer reviewers may or may not use these tools in their work and how they should disclose such use if they do, with many publishers enacting such policies. However, little attention has been paid to whether and how scholarly publishers disclose their own use of AI. This can include using AI in their publishing workflows, such as copy editing and image creation, but extends beyond as well. News items have reported on multi-million dollar deals publishers have made with tech companies to license their content to train AI tools or how scholarly publishers are creating their own AI tools based on their corpus of content. This presentation seeks to bring more attention to this issue by sharing the results of a content analysis of the largest scholarly publishers’ websites as well as the websites of their top journals. The analysis looked for publicly available language provided by the publishers about how they use AI and then analyzed the content through a lens of performative disclosure vs. meaningful disclosure. The presentation will also discuss how this issue affects library publishing programs and best practices that libraries should consider when deciding whether they need their own disclosure policies or how they should advise their editors and other participants. Even those who are not actively using AI are still part of the scholarly communications ecosystem, which means they are likely affected indirectly by AI.
