Oxford updates AI disclosure guidelines for academic authors
Oxford University Press (OUP) has updated its author and editor guidelines on the responsible, appropriate and transparent use of artificial intelligence, following research that found widespread uncertainty among academics over when AI use should be disclosed.
The publisher’s global survey of more than 2,600 researchers found that nearly three-quarters (74%) were unclear about what use of AI they should disclose when preparing academic work, despite strong support among researchers for greater transparency.
The news comes on the same day that Springer Nature published its own framework around AI use by authors.
The findings also highlighted concerns about how disclosure could affect perceptions of research. Some 44% of respondents said they were concerned that disclosing their use of AI would negatively affect how others viewed their work, while only 36% said they currently recorded their use of AI during the research process.
OUP says the findings point to a growing gap between the use of AI in academic research and its disclosure, and demonstrate the need for clearer and more consistent policies.
The updated guidelines come as efforts continue to develop a global reporting standard for AI disclosure in research. OUP says its approach is intended to give researchers greater confidence about recording and disclosing their use of AI tools.
The updated guidance applies to authors of both books and journal articles and focuses on three main areas:
- Authors are expected to disclose any significant use of AI tools and be transparent about how those technologies have contributed to their work.
- The guidelines also emphasise human oversight and accountability, with authors remaining responsible for reviewing and verifying AI-generated content used in books and journal articles.
- OUP has additionally strengthened guidance around intellectual property and confidentiality, advising researchers to take particular care when using AI tools with unpublished research, copyrighted material or sensitive content.
David Clark, Managing Director of OUP’s Academic division, said: “Our research into AI disclosure follows our 2024 survey, which revealed a wide range of initial perceptions towards AI. Two years on, AI use is more widespread, but there remain major differences in how researchers are adopting, applying, and disclosing these tools.
“At OUP, we’ve developed AI guidelines for authors with the view that the scholarship we publish must remain valued and protected. We regularly review these against industry guidelines and consider any feedback to ensure clear guidance for authors. We continue to prioritise working with the research community to develop clear publishing standards that support readers and authors in an AI-enabled world.”
Key findings of the survey include:
AI is helping researchers:
- 64% of researchers report AI as beneficial to their research, with over two-thirds (68%) using AI chatbots on the open web for their research, and just under half (49%) using machine translation.
- AI tools are mostly commonly used for discovering existing research (55%), summarising existing research (46%), and editing the write-up of research (45%). Many (68%) agree that AI tools can reduce accessibility barriers, too.
- However, 35% of those in the arts and humanities never use AI tools in their research, as opposed to 15% and 18% for STM and the social sciences, respectively.
Trust in AI is low:
- Only 20% trust AI-generated outputs. 37% of researchers who haven’t used AI for their research report concerns about accuracy, bias, and trustworthiness, and only 26% find AI-generated literature summaries to be accurate.
- However, 83% of those surveyed have established practices for checking citation sources generated by AI, and 80% always cross-check AI outputs for reliability.