ReVoicer: Conversational Voice Annotation for Human-Centered, LLM-Assisted Peer Review

Abstract

We present ReVoicer, a prototype system that supports peer reviewers by letting them converse with a paper as they read it. The reviewer highlights a passage and speaks (or types) a train-of-thought comment. A large language model then cleans the comment using the surrounding prose as context, tags it by comment type, and anchors it to the passage. After the reviewer finishes reading, ReVoicer checks the accumulated notes against a venue-specific rubric and reports coverage gaps to assist with further reflection. Then ReVoicer drafts a review composed from the reviewer's comments, written to a style guide distilled from the reviewer's past reviews. We describe the system's design rationale and implementation, and we outline plans for future evaluations.

IEEE ISMAR 2026 Alt'ISMAR workshop