Artificial Intelligence in Addiction Scholarship: Transparency, Accountability, and Disclosure in Research and Publishing

Authors:
Publisher website
Google Scholar
Abstract
Artificial intelligence (AI) is embedded in addiction scholarship, not only as a methodological tool in research but also as a tool for manuscript preparation. In addiction research and clinical care, AI is already used to analyze data and support clinical decision-making. In parallel, AI tools are increasingly used to summarize literature, revise prose, generate outlines, draft text, and assist with interpretation and presentation of findings. These developments create an urgent need for addiction journals to clarify how AI use should be governed and disclosed. In this editorial, we argue for a principled, proportionate framework for AI disclosure grounded in materiality. We distinguish AI used in the conduct of research from AI used in manuscript preparation and propose a practical taxonomy of assistive, intermediate, and generative AI mapped onto 6 escalating levels of involvement. We identify intermediate AI as a key governance challenge because it may appear to provide editorial assistance while materially shaping scholarly content. We argue that disclosure should be required when AI materially contributes to the conduct of research or to the intellectual content, interpretation, argument, or presentation of a manuscript, while routine low-risk assistive uses should not require disclosure. Transparent governance, organized around materiality rather than specific technologies, is the appropriate response to AI in addiction publishing.
Funding Information
  • National Institute on Drug Abuse (UG1DA049444)
  • U.S. Department of Veterans Affairs (CIN 13-414)

This publication has 36 references indexed in Scilit: