From Shadow AI to Governed AI Review: A Commentary on “Fighting Fire with Fire: Infusing AI into Peer Review to Sustain Quality Scholarship”

I strongly support the initiative advanced in “Fighting Fire with Fire: Infusing AI into Peer Review to Sustain Quality Scholarship” (Bhargava et al. 2026). The question facing journals is no longer whether artificial intelligence (AI) will enter the peer review process, but whether its use will remain informal, hidden, and unevenly governed, or become transparent, auditable, and institutionally accountable. The paper is persuasive in identifying the problem of “shadow AI”: Reviewers may already be using AI tools privately, with unknown prompts, unknown data protections, and unknown effects on reviewer independence. A journal-managed AI review process therefore offers an important opportunity not to replace human judgment, but to bring AI use into a governed workflow where its influence can be observed, evaluated, and improved.

At the same time, I would encourage journal editors and the scholarly community to give greater attention to several implementation risks. First, AI hallucination remains a serious concern. Human involvement should therefore be designed not only as final oversight, but as an accountability mechanism that actively verifies, contextualizes, and, when necessary, rejects AI-generated comments. Second, the process should consider authors’ perceptions of procedural dignity. If authors are asked to respond first to an AI-generated review before any human engagement, some may feel that their scholarly work has been evaluated by an impersonal system rather than by the academic community. This does not undermine the proposal, but it suggests the need for careful framing: AI feedback should be presented as a diagnostic and developmental input and not as a substitute for scholarly judgment. Finally, although I agree that AI should initially be restricted from publication recommendations and broader value judgments, the boundary between structured assessment and value judgment may evolve as AI capabilities develop. The community should therefore view the proposed workflow as an important starting point for continued experimentation and refinement as AI capabilities, reviewer practices, and scholarly norms continue to evolve.