STM, a global trade association for academic and professional publishers, has released a draft report titled 'Recommendations for a Classification of AI Use in Academic Manuscript Preparation', developed by its Task & Finish Group on AI Labelling Terminology. The draft is now open for public consultation.
As generative AI tools increasingly support tasks such as writing, editing, and generating visual content, the scope of machine-assisted manuscript preparation has expanded beyond traditional forms of human-disclosed support, such as language editing. Despite this evolution, existing publisher guidelines have not adapted at the same pace, creating uncertainty across the scholarly communication ecosystem.
The absence of consistent standards has raised concerns regarding transparency and trust. Authors face uncertainty over what forms of AI usage require disclosure, peer reviewers encounter ambiguity around acceptable AI contributions, and readers lack clear indicators to distinguish between human- and machine-generated content.
To address these gaps, the draft classification introduces a structured framework that publishers can adapt to their editorial processes. It aims to clarify the role of AI in manuscript development while reinforcing transparency and research integrity.
The draft builds on STM’s earlier 2023 guidance, offering a practical tool to support the development of policies that align with shared principles of accountability, transparency, and integrity across research disciplines.
Click here to read the original press release.
