Home/Insights/Technology innovations in research and publishing
Technology innovations in research and publishing
Generative AI is now present in writing, screening, peer review, integrity checking and production. These pages set out where it helps, where it is prohibited, and where the rules are unsettled.
What is covered
Best practice for authors writing with AI assistance. The role AI already plays in scholarly publishing, much of it invisible to authors. The opportunities and challenges taken together. And the ethical questions — accountability, transparency, equity and confidentiality — which are further from settled than the technical ones.
The positions that are settled
An AI tool cannot be an author: ICMJE and COPE are both explicit, because authorship requires accountability a tool cannot carry. Use must be disclosed. Authors remain fully responsible for AI-assisted text, including fabricated citations.
Uploading a manuscript under review to a general-purpose AI tool is a confidentiality breach, and several funders and publishers prohibit it outright.
The positions that are not
Whether and how publishers should disclose their own use of automation. What appeal exists when an automated system influences an editorial decision. How to handle AI-text detection, given that detectors are unreliable and their false positives fall disproportionately on authors writing in a second language.
In short
What to take from this.
- Settled: AI cannot be an author, use must be disclosed, authors stay responsible.
- Settled: do not put manuscripts under review into general AI tools.
- Unsettled: publisher disclosure, appeal rights, and detection.
Sources
- ICMJE and COPE positions on AI and authorship
- WAME recommendations on chatbots in scholarly manuscripts
Related