True or False: Publishers’ policies around ChatGPT and other AI tools

by Mélodie Garnier

As ChatGPT gains popularity with writers worldwide, academic publishers are having to define their stance on the use of AI tools.
Here’s the low-down on their official policies.



Publishers ban the use of ChatGPT in the preparation of manuscripts

It depends
Perhaps surprisingly, most academic publishers do allow authors to use ChatGPT and other tools when writing papers. The policies in place are more to do with ensuring responsible use. AI tools such as ChatGPT are here to stay, and as language models become better and better (see GPT-4), the output may soon be indistinguishable from that of humans anyway. Banning it outright may be superfluous.
Still, some publishers do ban it unless explicit permission is granted from editors, such as Science. Always make sure to read a journal’s policy on AI tools before submitting a manuscript.


If ChatGPT is allowed, I can copy-paste AI-generated content into my text

False
If ChatGPT is allowed, don’t assume you have a free pass to use it as you please. All publishers agree that AI tools should be used carefully and responsibly.
For Elsevier, “[they] should only be used to improve readability and language of the work. Applying the technology should be done with human oversight and control and authors should carefully review and edit the result, because AI can generate authoritative-sounding output that can be incorrect, incomplete or biased.”
Taylor & Francis makes clear that “authors are accountable for the originality, validity and integrity of the content of their submissions”. When using AI tools, “authors are expected to do so responsibly and in accordance with our editorial policies on authorship and principles of publishing ethics.”


If I use ChatGPT when writing my paper, I should credit it as an author

False
There are clear rules on how to acknowledge AI tools in your scientific paper. You may be tempted to list ChatGPT next to your name to give it full credit. But all major publishers now say AI tools cannot be acknowledged as a co-author. Nature justifies this by stating that “any attribution of authorship carries with it accountability for the work, and AI tools cannot take such responsibility.” A machine cannot justify this or that method or this or that statement. It has simply digested and produced content from the internet.


If I use ChatGPT when writing my paper, I should credit it in the methodology, references, or acknowledgements

True
Failing to do so technically constitutes plagiarism, which is a breach of academic integrity and a severe offense. SAGE even requires to indicate “which model was used and for what purpose”, as well as “limitations of language models [...] including the potential for bias, errors, and gaps in knowledge.” According to Elsevier, “declaring the use of these technologies supports transparency and trust between authors, readers, reviewers, editors and contributors and facilitates compliance with the terms of use of the relevant tool or technology.”