InfoResearchPeer-reviewedLLM-specific
From Policy to Architecture: A Review of Confidential, Human‐Accountable AI Support for Scholarly Peer Review
- Published
- Record updated
Summary
A structured narrative review of 59 sources (48 cited) examines how editorial and publisher policies on generative AI in peer review address confidentiality, accountability and disclosure. The authors find these policies remain predominantly behavioural, with no complete technical specification for auditability or enforcement, and they identify 12 architecture-level design principles, such as local-first processing and auditable disclosure trails, as a basis for reviewer-support tools.
Mitigation
Publishers, journals and editorial organisations should move beyond behavioural guidance to specify technical expectations for reviewer-support tools; empirical validation remains a priority.