Toward Structured Transparency: A Governance Framework for AI Use in Biomedical Publishing

Area of Science:

• Biomedical research and publishing
• Artificial intelligence in science
• Scientific communication

Background:

• Rapid adoption of generative AI and LLMs transforms biomedical research and publishing.
• Existing policies from ICMJE and COPE restrict AI authorship but lack operational detail.
• Journal instructions for authors need enhancement for consistent AI principle application.

Purpose of the Study:

• Critically examine AI policies in JKMS and KAMJE.
• Identify challenges in AI non-authorship, disclosure, and conflicts of interest.
• Propose a framework for structured transparency in AI use.

Main Methods:

• Review of current AI-related policies in JKMS and KAMJE.
• Analysis of practical limitations and emerging challenges.
• Development of a proposed governance framework for AI integration.

Main Results:

• Identified three major challenges: AI non-authorship limitations, inadequate AI disclosure, and AI-specific conflicts of interest.
• Current policies are insufficient for systematic documentation, evaluation, and verification of AI use.
• A shift from restrictive approaches to structured transparency is necessary.

Conclusions:

• Medical publishing requires a move towards structured transparency for AI integration.
• Proposed framework includes tiered disclosure, author accountability, and conflict-of-interest management.
• Prioritizing transparent AI governance strengthens research integrity and responsible AI adoption.