Use of generative AI tools
These policy guidelines cover the use of generative AI tools (including Large Language Models or LLMs) and AI-assisted technologies by authors, reviewers and editors. GEUS Bulletin will update these guidelines as these fast-changing technologies develop.
Use of AI by authors
Journal policy for the use of AI in manuscripts submitted to GEUS Bulletin is developed in line with advice from the Committee of Publishing Ethics (COPE).
Generative AI tools do not meet the journal criteria for authorship and so cannot be listed as a co-author of articles submitted to GEUS Bulletin.
Permitted use of AI tools by authors includes copyediting of human-generated text (i.e. text originally written by human co-authors) to refine, correct, edit or format manuscripts to improve spelling and grammar, readability and style. Such usage does not need to be declared in the submitted manuscript.
Where AI tools are used for any other purpose, this must be declared and clearly documented in a Methods section for Research, Data or Method articles or in a separate dedicated section in other article formats. Authors must declare which tool was used and how (including prompts given).
AI tools must not be used to autonomously generate or summarise manuscript text (inc. references), images or other graphical elements of the manuscript. AI tools must not be used to collect, refine, format, analyse or display data or code without full disclosure in the manuscript. This also applies to supplementary files.
It is the responsibility of all co-authors to declare the use of AI in preparation of the manuscript, as outlined here. All authors are expected to have reviewed and approved all components of the manuscript submitted to GEUS Bulletin and revisions submitted during the review process, acceptance and publication.
Authors are fully responsible for the content of their manuscript, even those parts produced by an AI tool, and are thus liable for any breach of publication ethics.
Use of AI by reviewers and editors
To maintain confidentiality throughout the review process, reviewers and editors are strictly prohibited from uploading any component of the manuscript under consideration or review comments or editorial comments to any online, third-party AI tool, such as ChatGPT or similar.
The only type of AI tool permitted for use by reviewers and editors are those hosted on on-premises servers at the reviewer’s or editor’s affiliated institution, where the institution can guarantee that uploaded content has not and will not ever be shared online or sent to third parties.
AI must not be used to generate or summarise comments by reviewers or editors.
Permitted use of AI tools by editors and reviewers are limited to the following:
- Checks for spelling and grammar in programmes such as Microsoft Office suite, which makes use of AI under certain licenses.
- copyediting of review or editorial comments to improve spelling, grammar, readability and style of human-generated text (i.e. text originally written by human reviewers).
