Analysis of generative artificial intelligence tools usage policies provided to authors by journals publishing aquaculture research

Abstract

Generative artificial intelligence (GenAI) is influencing how research is conducted and reported, however there is limited agreement on what constitutes acceptable use in scholarly publishing. Publishers contribute to shaping norms around GenAI, yet journal policies vary. This study compared GenAI policies across aquaculture journals, especially those publishing research on aquatic animal diseases, to raise awareness to variability in acceptable use. GenAI policies for nine journals were analysed within four domains: Policy Status, Scope and Permissions, Author Responsibilities, and Accountability. All journals explicitly or implicitly permitted GenAI use in manuscript preparation, although detailed guidance was often lacking for specific tasks. Considerable variation existed between journals concerning the creation of figures, from permitted (1/9) and implicitly permitted (3/9) to prohibited (3/9). All journals required disclosure of GenAI use and generally provided instructions, although in some cases certain tasks did not require disclosure including use for grammar and spelling (4/9 journals). Authors were explicitly responsible for accuracy (4/9) and originality (6/9) of GenAI-assisted content, while most journals provided little explicit guidance on legalities of GenAI use or for issues pertaining to data privacy, confidentiality, intellectual property rights, and data ownership, with 4/9 journals not providing any guidance on these matters. Most journals (7/9) explicitly prohibited GenAI authorship. Actions that may help protect research integrity and trust include clearer journal guidance for GenAI use, standardised disclosure statements, and support for researcher training. This is the first study to highlight gaps and variations in aquaculture journal policies to GenAI, which supports authors and contributes to the ongoing debate around acceptable practices.