This blog was co-authored by Gráinne McNamara (Research Integrity and Publication Ethics Manager), Teresa Mathews (Publication Manager and Researcher Engagement and Recognition Liaison), and Beth Bayley (Open Science Manager).
The theme of Peer Review Week (PRW) this year is ‘Rethinking Peer Review in the AI Era’ which prompts an important reflection on responsibility and accountability when it comes to use of AI in peer review and the crucial role of human oversight as we engage with emerging technologies.
AI can support streamlining of some peer review processes, but can also introduce new risks: fabricated reviews, reviewer identity fraud, and over-reliance on automated assessments. For Peer Review Week this year, we are exploring the duty we have to encourage responsible innovation in peer review and, specifically, how co-review can support editors and reviewers in ensuring the integrity of our publications.
Co-reviewing offers early-career researchers a hands-on apprenticeship in the ethics, rigor, and responsibilities of peer review. In a time when AI tools can generate convincing but potentially misleading content, this human-centered mentorship becomes even more critical. Karger introduced co-reviewing in a pilot program across 14 journals in April 2025 integrating co-reviewers into the peer review process and enabling them to easily receive recognition of their participation. Following this successful pilot, in July 2025, we expanded co-review to all journals to support reviewers and Editors in training the next generation of reviewers. In the first month of the global rollout, co-reviewer inclusion requests averaged 2.3 per day with over 75% of invited co-reviewers confirmed their willingness to participate. This early success reinforces the importance of practical training for early career reviewers, complementing the structured training modules we provide.
This initative has been welcomed by our Editorial Boards as a way of supporting early career researchers. Speaking with Dr. David Vaudry, Editor in Chief of Neuroendocrinology, he shared that he “strongly encourage[s] co-review to train the next generation of reviewers. In general, I think earlier career researchers have more time to thoroughly read the manuscripts. For me, there is also less risk to have a conflict of interest because they don’t know all the community already. However, one might sometimes feel uncertain about criticizing articles from eminent authors in their field. Therefore, if they can be supported by a senior reviewer through co-review, for us it is confirmation that the work has been well done”.
Co-review can support the professional development of early career researchers, beyond receiving recognition of their contribution. Dr. Vaudry echoed this sentiment in our conversation sharing that acting as a co-reviewer can be “an excellent opportunity to develop critical reading of articles and better understand the strengths and weakness of manuscripts… It’s also an opportunity for them to progress in the writing of their own articles because they become more critical of their work by putting themselves in the shoes of a reviewer when they write their own manuscripts”
Looking ahead for innovation in peer review, Karger will be rolling out a pilot in Transparent Peer Review, whereby an article’s peer review history will be published alongside it. By openly sharing reviewer comments, author responses, and editorial decisions, we can foster more trust and accountability in scholarly publishing—values that algorithms cannot uphold. Similar to co-review, transparent peer review provides real learning opportunities: Openly sharing this human expertise also benefits other researchers, who can learn from the unique insights in the reviewers’ comments and their exchanges with the authors.
As in many other areas in scholarly communication, AI is a mighty tool, but it should supplement, not replace, human judgement and should be used responsibly to the extent it truly benefits us.
Acknowledgements
This draft was prepared with assistance from Microsoft co-pilot with extensive editing. Included data was calculated by the authors without AI assistance.
