Ethical Guidelines for Artificial Intelligence Use in Scholarly Publishing for Annals of Geriatric Medicine and Research: A Framework for Integrity

The rapid advancement of generative artificial intelligence (AI) is transforming the academic publishing landscape. A growing number of specialized generative AI tools offer efficiency in areas such as data synthesis, literature review, and manuscript preparation.¹) However, they also introduce challenges regarding ethical integrity, intellectual property, data privacy, and the risk of AI "hallucinations" in which generated content may appear plausible but is factually incorrect.²) Given the clinical complexity and multidisciplinary nature of geriatrics, human clinical insight remains indispensable and cannot be substituted by generative AI. To uphold the quality, credibility, and scientific rigor of the Annals of Geriatric Medicine and Research (AGMR), the journal has established clear guidelines governing the responsible use of AI by authors, reviewers, and editors. These policies on AI are aligned with the ethical standards and recommendations of the Committee on Publication Ethics (COPE) and the International Committee of Medical Journal Editors (ICMJE), both of which emphasize that responsibility and accountability for published work must remain with human authors.^3,4)

AI Authorship

While authors may acquire information and assistance from generative AI, such tools cannot be listed as authors. Authorship is strictly limited to human individuals or groups who have made substantial intellectual contributions and take full accountability for the work. According to ICMJE, authorship is given to those who can take responsibility for the accuracy or integrity of the work, indicating that AI tools do not qualify for authorship.⁴)

If generative AI is used, appropriate disclosure is essential to ensure transparency. AGMR requires authors to clearly state the extent and purpose of AI use both in the manuscript (in an appropriate section) and in the cover letter.⁴) Authors should specify the name of the AI tool, its version, the manufacturer, and how it was used. Disclosure is not required for AI-assisted tools used solely for routine language editing, such as grammar, spelling, or reference formatting checks. As AGMR provides a platform for researchers from emerging countries, many of which are non-English speaking,⁵) the use of AI-assisted tools for refining language is permitted to bridge academic gaps, enhance readability, and enrich the quality of scholarly papers.

Regarding visual data, AGMR does not permit images and figures produced by generative AI. AI-generated images raise significant concerns regarding scientific accuracy, and may lead to clinical misinformation, as well as copyright and legal issues related to infringement of AI training data. Nevertheless, if AI-assisted tools are used as part of the research, authors must clearly identify the tool used directly in the figure or legend. In addition, authors must describe the type of AI-assisted tool used in the method section and explain how it was applied.³)

Authors must also be cautious regarding the phenomenon of AI "hallucinations," whereby generative AI systems may produce content that appears accurate but is, in fact, incorrect or fabricated.⁶) The inclusion of erroneous information, such as nonexistent references, inaccurate data, or misleading interpretations, can seriously compromise the scientific integrity of the journal. Accordingly, authors bear full responsibility for verifying all content generated or assisted by AI tools and must always double-check all information provided by generative AI.

Use of AI Tools in the Peer Review Process

High-quality peer review relies on rigorous, independent evaluation of human experts. AI tools cannot replace this process, as they may introduce bias, inaccuracies, or misleading information. Peer reviews must reflect the reviewers’ own scholarly judgment, and reviewers retain full responsibility and accountability for their comments.

Confidentiality is a fundamental principle of peer review. Reviewers must recognize that the manuscript under review is an original and valuable work into which the author has invested significant time. If a manuscript is uploaded to a generative AI platform, the security and privacy of the content cannot be guaranteed, possibly exposing the author’s intellectual property.⁴) Therefore, AGMR requires reviewers to rely on their clinical and academic expertise and refrain from using generative AI when reviewing manuscripts. However, if a reviewer uses an AI tool as a resource in a way that does not breach these privacy standards, the extent and purpose of such use must be disclosed in the “to the editor” section for the sake of transparency.

AI will continue to evolve and shape scholarly publishing, presenting both opportunities and ethical challenges. AGMR is committed to preserving scientific integrity by proactively and responsibly navigating the ethical challenges of AI. We will regularly update our policies to ensure that our standards remain aligned with those of COPE and the ICMJE.