Editors’ Note: Today’s post is by Kari D. Weaver, Bert Seghers, Kiera McNeice, Mike Perkins, Sergio Santamarina, Lorelie Lingard, Natalia Tsybuliak, and Felix Dijstal. Kari is Program Manager, Artificial Intelligence and Machine Learning, Ontario Council of University Libraries. Bert is Secretary, Flemish Commission for Research Integrity and President, ENRIO. Kiera is Research Transparency Manager, Cambridge University Press & Assessment. Mike is Head of the Centre for Research & Innovation, British University Vietnam. Sergio is Library Coordinator, National University of José C. Paz. Lorelei is Distinguished University Professor, Western University. Natalia is Associate Professor, Berdyansk State Pedagogical University. And Felix is Science Officer, International Science Council.
Research Integrity in the Era of AI
Following the launch of ChatGPT, researchers quickly recognized the potential for using generative artificial intelligence (GenAI) tools in their work, as well as the potential academic integrity concerns this may give rise to. In this context, the longstanding expectations that researchers acknowledge work that is not their own, and describe their methods transparently, have become more relevant than ever. Researchers should be as transparent about their use of artificial intelligence (AI) as about any other contributions to scholarly outputs — this is a matter of research integrity comparable to disclosure of funding sources, conflicts of interest, and acknowledgements.
The Need for Clarity and Consistency
The rapid integration of AI into research and scholarly publishing processes has created an urgent need for clear, consistent, and community-appropriate approaches to disclosure. While there are many guidelines for the responsible use of AI in research, and technical initiatives are advancing machine-readable infrastructures for automated tracking of AI contributions, a critical gap remains: there is no shared standard for how researchers should disclose the ways AI use has contributed to scholarly outputs.
This creates confusion for authors, reviewers, editors, and institutions alike, and risks undermining the transparency, reproducibility, and integrity of research. Further complicating matters, research has highlighted structural and behavioral reasons for avoiding disclosure; for example, studies show that disclosing the use of GenAI tools can negatively affect trust in scholarly outputs.
Addressing the AI Disclosure Gap
The World Conferences on Research Integrity (WCRI) have convened a Focus Track to address this gap in clarity and standardization of AI disclosure. While this alone will not solve the broader challenges surrounding perceptions of AI use in research, we believe a harmonized approach to AI disclosure is a necessary foundational step in service of cultural change towards genuine reflection, mindful documentation, and meaningful transparency regarding AI use.
Recent evidence underscores the urgency of this work. Results of a 2026 survey show that AI use is already embedded in everyday research practices, with a majority of researchers using it to support activities ranging from writing and translation to data analysis and peer review. However, in the same survey, around a third of researchers reported they have never disclosed their use of AI when publishing their work. Even if researchers have the best intentions regarding transparency and integrity, closing this “disclosure gap” will require a shared understanding and clarity about what meaningful disclosure should entail. In the absence of wide-spread standards, researchers, publishers, and institutions are struggling to navigate inconsistent norms and expectations regarding how AI use should be disclosed in scholarly work.
A Global Consultation
To work towards a global reporting standard for AI disclosure, the WCRI Focus Track has brought together a core team of experts and originators of existing AI disclosure frameworks. We accepted the invitation to help support this trajectory as the “core team” of this focus track. We will process the results of the consultation rounds and develop the materials for the next. With the valued support of the International Science Council (ISC), Committee on Publication Ethics (COPE), International Association of Scientific, Technical, and Medical Publishers (STM), and the Global Young Academy, we reach out to the broader research community, with an invitation to contribute to three rounds of consultation, focusing (among other things) on:
- Dec 2025–Feb 2026: Mapping community needs and views on disclosure structure
- July 2026–Oct 2026: Identifying and proposing a taxonomy for disclosure components
- Late 2026-Early 2027: Refining a draft “Vancouver Standard” for AI disclosure
Initial Insights
We have completed our analysis of responses to the first round of consultation, where results highlighted both the complexity of the landscape of AI use and disclosure, and a strong appetite for clarity. Key points emphasized by contributors were the need to support transparency without imposing undue reporting burdens, and ensure suitability across disciplines. Common themes included questions of granularity (what level of detail is appropriate), consistency (how to align expectations and practices across fields), and interoperability with existing and future processes and infrastructure. There was also significant discussion about the purposes of disclosure — whether it should be primarily about accountability, reproducibility, ethical transparency, or all of these — and how this purpose should inform the nature of any standards.
Consensus emerged around the view that while GenAI tools may represent a step change in the way that research is conducted and reported, long-established principles and frameworks of research integrity, transparency, and accountability apply to the use of GenAI tools as much as to the use of any other sophisticated software in the research lifecycle. From this perspective, the use of AI in and of itself should not be stigmatized.
There was also general agreement that disclosure of the use of AI in research processes should focus on the integrity and credibility of research methods and conclusions. Human accountability, responsibility, and oversight remain essential to the responsible use of AI in research. Disclosures should provide enough information for a disciplinary peer, whether a reviewer or reader, to reasonably assess whether AI contributions may have affected the reliability of the reported work.
However, complicated questions remain, including around where and how AI use should be disclosed in the main manuscript content or as a separate statement; these will be explored in further rounds of consultation.
Ongoing Engagement
Following the first consultation round and further discussions within the core team, we continue to work towards an AI disclosure standard that will encourage genuine reflexivity, mindful documentation, and meaningful disclosure about the use of AI tools in the research life cycle.
The core team held two dedicated sessions May 4-5 at the 2026 World Conference to engage directly with research integrity experts from all sectors. These focused on the question of thresholds for when AI disclosure becomes necessary, a proposed taxonomy of controlled categories to use as a framework for disclosure, and current and potential future applications of AI across the research life cycle.
In early July, our second global consultation round will open for responses. Details will be published on the ISC website, and we encourage participation and input from all stakeholders across the scholarly communications landscape to help shape the ongoing development of this work.
We are aware that for these efforts to succeed, it is essential that our work is globally inclusive and genuinely supportive of the needs of all research communities. As well as soliciting views on standards for disclosing the use of AI in research, the core team welcomes any questions, suggestions, or concerns about the approach we are taking, as well as expressions of interest in other ways to participate and contribute, via office@enrio.eu.
