The rapid adoption of generative artificial intelligence tools in academic writing has prompted major scholarly publishers to establish guidelines governing their use. This paper provides a comprehensive analysis of AI-assistance policies from eleven major publishers: ACM, Cambridge University Press, Elsevier, IEEE, Nature Portfolio, Oxford University Press, SAGE Publishing, Taylor & Francis, and Wiley. Through systematic comparison, we identify a strong consensus on core principles including the prohibition of AI systems as authors, the requirement for transparency through disclosure statements, and the necessity of human accountability for published work. However, notable differences emerge in policies regarding AI use in peer review, image generation, and the level of detail required in disclosure statements. We document that all examined publishers converge on treating AI as an assistive tool requiring human oversight rather than an autonomous contributor to scholarship. The analysis reveals that while the fundamental ethical framework is shared across publishers, implementation details vary considerably, reflecting different institutional priorities and disciplinary considerations. These findings provide clarity for researchers navigating the evolving landscape of AIassisted academic writing and contribute to the broader discourse on maintaining scholarly integrity in an era of generative AI.
Rachel So. Publisher Guidelines on AI-Assisted Scholarly Writing: A Comparative Analysis. Project Rachel, November 2025. https://doi.org/10.71775/kth.57ph6-8fr16