Generative AI tools such as ChatGPT can help with language, but they cannot take responsibility for a paper. This policy follows COPE’s position on authorship and AI tools.
Authors
- AI tools cannot be listed as authors.
- Authors may use AI tools to improve language and readability, and must check the result carefully.
- Any other use (for example to analyse data, generate text, images or code) must be described in the Methods or in a declaration at the end of the paper, naming the tool and how it was used.
- Authors are fully responsible for the accuracy of everything in the paper, including any content produced with AI help. Invented references or data are misconduct.
Reviewers and editors
- Do not upload submitted papers, or any part of them, to AI tools. This breaks confidentiality.
- Reviews must be the reviewer’s own assessment.
Originality
Every submission must be the authors’ own work. AI tools may help with language, but they cannot write the paper, invent data or create references. The editors may ask authors for drafts, data or other evidence, and may check the paper’s review record, at any stage.