Assessing AI integration in psychiatry research: Evaluating policies across leading journals

Introduction/Objectives: By refining systematic reviews, enhancing data analysis, and discovering clinical applications, AI is revolutionizing psychiatry and mental health research. However, the rise and mounting reliability of AI in research bodes significant concerns about transparency, ethics, and reproducibility. With the advent of generative AI, human prompting is no longer necessary in creating, reviewing, and drawing conclusions from data. This study examines how leading psychiatry and mental health journals address these challenges through their author instructions and editorial policies.

Methods: A cross-sectional review of the top 100 peer-reviewed psychiatry and mental health journals, ranked by the 2023 SCImago SJR indicator, was conducted. Data from each journal's "Instructions for Authors" were extracted to assess AI-related policies, including authorship criteria, guidelines for reporting AI use, and the acceptance of AI-generated content (e.g., manuscript preparation and image generation). Correlational analyses were then performed to examine associations between these policies and journal characteristics.

Results: 100 journals were assessed and 91% of them included AI use in their author instructions. Most prohibited AI authorship, but required disclosure of AI involvement in submissions. AI-generated content was allowed by 22% of journals, and 13% permitted AI-generated images. Journals with higher impact factors were more likely to have detailed AI policies. However, significant gaps in standardization and guidance persist.

Conclusions: While many psychiatry and mental health journals recognize AI's growing role in research, few have implemented specific reporting guidelines (RGs) for its use. This glaring absence of standardized guidelines threatens both the transparency and the integrity of AI-driven research. Without comprehensive regulations, the risks of unethical practices and irreproducible results loom large. To safeguard ethical and high-quality research in this rapidly evolving era, the adoption of robust AI guidelines is urgently needed.