Conducting Eating Disorder Research in the Era of Generative AI: Researcher Perspectives and Guidelines From the International Journal of Eating Disorders.

Objectives: Generative Artificial Intelligence (AI) could transform how science is conducted, supporting researchers with writing, coding, peer review, and evidence synthesis. However, it is not yet known how eating disorder researchers utilize generative AI, and uncertainty remains regarding its safe, ethical, and transparent use. The Executive Committee of the International Journal of Eating Disorders disseminated a survey for eating disorder researchers investigating their practices and perspectives on generative AI, with the goal of informing guidelines on appropriate AI use for authors, reviewers, and editors.

Method: A survey was distributed globally via eating disorder organizations, professional networks, and individual researchers. Researchers (N = 158) of various career stages completed the survey.

Results: Nearly three-quarters (70%) reported using generative AI for research, most commonly for proofreading written work or coding support. Nine in 10 took steps to verify AI-generated output, and 1 in 3 disclosed their use of AI. Only 21% reported using AI for peer review, typically in a limited capacity (e.g., proofreading), and always with full human oversight. Authors were comfortable for editors to use AI to support administrative tasks (i.e., selecting reviewers, detecting plagiarism). However, many participants acknowledged key drawbacks of generative AI, including concerns about inaccurate outputs, ethical issues such as plagiarism, the potential for reduced critical thinking, and anticipated negative impacts on the future of eating disorder research.

Conclusion: These insights informed the development of field-specific guidelines to support authors, reviewers, and editors in the appropriate use of generative AI in eating disorder research and publishing.