Navigating the AI Revolution: A New Chapter for the Diabetes and Metabolism Journal
Summary: Published March 1, 2026 in Diabetes & Metabolism Journal (Vol. 50, No. 2, pp. 253–254; Korean Diabetes Association), this editorial by incoming editor-in-chief Junghyun Noh (Division of Endocrinology and Metabolism, Inje University Ilsan Paik Hospital, Goyang, Korea) announces the journal’s evolving policy framework for artificial intelligence use in scientific research and manuscript preparation. The editorial follows a broader trend across academic publishing: an Elsevier survey of 3,234 researchers from 113 countries found that AI tool use in research rose from 37% in 2024 to 58% in 2025, with expectations of continued growth. Dr. Noh identifies four core concerns the journal is addressing: (1) scientific integrity and originality, as AI-assisted text may contain inaccurate interpretations or fabricated references; (2) data and image fabrication risks, as advanced AI systems can produce synthetic datasets and manipulated figures that are difficult to distinguish from genuine outputs; (3) authorship and contributorship ambiguity, requiring clearer disclosure norms; and (4) the substitution risk — AI-generated text may appear fluent and grammatically correct while lacking the scientific depth, critical analysis, and domain expertise that peer-reviewed work demands. DMJ’s immediate policy response requires authors to disclose all AI tool use during manuscript preparation, data analysis, or figure development — specifying tools, applications, and confirmation that outputs have been verified for accuracy. This disclosure must appear in the Methods section or as a dedicated AI Assistance Statement and will be published with the article. The journal is also evaluating AI-based tools for editorial screening of data and image manipulation, training editorial staff, and signalling openness to well-conducted AI-methods studies in dedicated future formats.
Key Highlights:
- New policy (immediate effect): all AI tool use during manuscript preparation, data analysis, or figure development must be explicitly disclosed, with tools named and AI-assisted content confirmed as author-verified for accuracy
- Elsevier 2025 survey context: AI use among researchers rose from 37% to 58% in one year across 113 countries; most anticipate greater efficiency gains ahead — underscoring the urgency of journal-level governance frameworks
- Four key integrity risks identified: fabricated or inaccurate AI-generated text; synthetic datasets and manipulated figures; authorship ambiguity; and substitution of AI fluency for genuine scientific judgment and domain expertise
- Editorial safeguards in development: evaluation of AI-detection tools, particularly for image and data manipulation screening; enhanced training for editors and reviewers; engagement with cross-journal AI ethics initiatives
- Openness to AI-methods research: if a sufficient body of rigorous, transparent AI-methods studies emerges, DMJ may create a dedicated section — focusing on drug development, risk assessment, predictive modeling, and precision medicine in diabetes and metabolism
- Authorship principle maintained: AI tools may not be listed as authors; the corresponding human authors bear full responsibility for the accuracy, integrity, and originality of all submitted content
Keywords: AI diabetes research publishing, artificial intelligence wound care research, scientific publishing integrity, AI manuscript disclosure, diabetes metabolism journal, LLM medical research
Junghyun Noh