Author Guideline:

The submission and publication of content created by artificial intelligence, language models, machine learning, or similar technologies is discouraged, unless part of formal research design or methods, and is not permitted without clear description of the content that was created and the name of the model or tool, version and extension numbers, and manufacturer. Authors must take responsibility for the integrity of the content generated by these models and tools.


Additional information for authors:

AI Used in Manuscript Preparation

When traditional and generative AI technologies are used to create, review, revise, or edit any of the content in a manuscript, authors should report in the Acknowledgment section the following:

Name of the AI software platform, program, or tool
Version and extension numbers
Manufacturer
Date(s) of use
A brief description of how the AI was used and on what portions of the manuscript or content
Confirmation that the author(s) take responsibility for the integrity of the content generated
Note this guidance does not apply to basic tools for checking grammar, spelling, references, and similar.

AI Used in Research

When AI (eg, large language model [LLM] or natural language processing [NLP], supervised or unsupervised machine learning [ML] for predictive/prescriptive or clustering tasks, chatbots, or similar other technologies) is used as part of a scientific study, authors should:

Follow relevant reporting guidelines for specific study designs when they exist and report each recommended guideline element with sufficient detail to enable reproducibility.
Avoid inclusion of identifiable patient information in text, tables, and figures.
Be aware of copyright and intellectual property concerns - if including content (text, images) generated by AI, and indicate rights or permissions to publish that content as determined by the AI service or owner.
Also address the following:

Methods Section

Include the study design and, if a relevant reporting guideline exists, indicate how it was followed, with sufficient detail to enable reproducibility.
Describe how AI was used for specific aspects of the study (eg, to generate or refine study hypotheses, assist in the generation of a list of adjustment variables, create graphs to show visual relationships).
For studies using LLMs, provide the name of the platform or program, tool, version, and manufacturer; specify dates and prompt(s) used and their sequence and any revisions to prompts in response to initial outputs.
For studies reporting ML and algorithm development, include details about data sets used for development, training, and validation. Clearly state if algorithms were trained and tested only on previously collected or existing data sets or if the study includes prospective deployment. Include the ML model and describe the variables and outcome(s) and selection of the fine-tuning parameters. Describe any assumptions involved (eg, log linearity, proportionality) and how these assumptions were tested.
Indicate the metric used to evaluate the performance of the algorithms, including bias, discrimination, calibration, reclassification, and others as appropriate.
Indicate the methods used to address missing data.
Indicate institutional review board/ethics review, approval, waiver, or exemption.
Describe methods or analyses included to address and manage AI-related methodologic bias and inaccuracy of AI-generated content.
Indicate, when appropriate, if sensitivity analyses were performed to explore the performance of the AI model in vulnerable or underrepresented subgroups.
Provide a data sharing statement, including if code will be shared.

Results Section

When reporting comparisons, provide performance assessments (eg, against standard of care), include effect sizes and measures of uncertainty (eg, 95% CIs) and other measurements such as likelihood ratios, and include information about performance errors, inaccurate or missing data, and sufficient detail for others to reproduce the findings.
Report the results of analyses to address methodologic bias and population representation.
If examples of generated text or content are included in tables or figures, be sure to indicate the source and licensing information, as noted above.

Discussion Section

Discuss the potential for AI-related bias and what was done to identify and mitigate such bias.
Discuss the potential for inaccuracy of AI-generated content and what was done to identify and manage this.
Discuss generalizability of findings across populations and results of analyses performed to explore the performance of the AI model in vulnerable or underrepresented subgroups.