The use of Generative artificial intelligence in research is becoming widespread enough that it is time to set clear rules to make it explicit. Many institutions, as well as major scientific publishers, establish their own principles (especially in the US, Canada, Great Britain…) but some common elements can be identified:

• Indicating the use of spelling or grammar checkers, as well as “reference management” applications, is not essential.
• All other uses must be specified. For example, “During the preparation of this (dissertation, report), the author used XXX (specify the exact model used)”, for example: Open AI (2025) Dall-E (version 3) https://open.ai.com/
• Reference standards (e.g., APA 7th edition) provide guidance on how to proceed. Let us recall that AI is not considered an author… but it must be cited in the references. It may be useful, depending on the topic, to indicate the “prompts” used to interact with the chatbot.
• A stricter approach is adopted by some universities (e.g., Zurich) where, for dissertation, a detailed signed declaration must be included in the appendix, specifying the precise use of AI and the portions of the text concerned.

These recommendations are sometimes included in universities’ AI Usage Charters (See Montpellier, Toulouse, Aix, etc.: https://www.observatoire-la-formation.univ-amu.fr).

Another humorous solution is to mention the use of generative AI in the acknowledgments, for example with a sentence such as: “The authors thank the software engineers in charge of Gemini, Copilot, etc. for their valuable help with documentation and the rephrasing of certain paragraphs.”

Some other complementary references:

• https://www.uzh.ch: recommendations on the use of AGI according to the University of Zurich.
• https://zenodo.org/records/13751602: how to cite sources in the AI era, University of Lorraine.
• https://www.uqam.ca.libguides.com: Chat GPT and AGI
• https://www.elsevier.com: on the use of AI for Elsevier.
• https://thesify.ai: for ethical use in academic papers