Use of generative AI
What authors declare when they use generative AI, and what reviewers and editors must never do with a manuscript.
- Applies to
- Authors, reviewers and editors
- Last updated
- 26 Oct 2026
- Questions
- [email protected]
In short
- AI tools cannot be authors.
- Say which tool you used and for what; the statement is printed in the article’s Declarations.
- Reviewers and editors never upload a manuscript or a report to an AI tool.
Authors
AI tools cannot be authors, because they cannot take responsibility for the work. Authors who used generative AI to write or prepare a manuscript say so when they submit, naming the tool and what it was used for, and describe it in the methods or the acknowledgements; the statement is printed with the article, in its Declarations, as “Use of generative AI”.
Authors remain responsible for the whole article, including any text, data or references a tool produced. Spelling and grammar checkers and reference managers need no statement.
Images and data
Images, figures and data made with generative AI are not accepted, unless AI is itself the subject of the research; then the methods say how they were made and the legend says so.
Reviewers and editors
Reviewers and editors keep manuscripts confidential: they do not upload a manuscript, a report or a decision letter to a generative AI tool. The review form asks every reviewer to confirm this.
If use was not declared
The ethics team handles it as set out in Publication ethics: the authors are asked to explain. Undeclared use found after publication may lead to a correction or, in serious cases, a retraction.