Notice on NLPCC AI-related Peer Review Studies
The 15th CCF International Conference on Natural Language Processing and Chinese Computing
Overview
NLPCC will conduct two AI-related peer review studies during this year's conference review process. Both studies are research activities intended to help the community better understand the appropriate use, benefits, limitations, and risks of AI in academic peer review. Because the two studies differ in their research goals and in the way AI is involved in the review process, NLPCC provides this summary notice to help authors and reviewers understand the distinction between them.
Participation in either study will not affect the review, discussion, scoring, ranking, or final acceptance decision of any paper. Authors may opt out of either study according to the mechanisms described below and in the corresponding formal notices.
The Two Studies
AI-assisted Peer Review Study
Focuses on how AI can support reviewers during the normal peer review process. The system used is ReviewCopilot, with model support provided by Xiaomi MiMo. The AI system is used only as an assistance tool and will not independently make review judgments, assign scores, or replace human reviewers.
Read the formal noticeAI Review Fairness Study
Focuses on how AI-generated reference information may affect human reviewers' scores, rationales, novelty judgments, and recommendations. Reviewers may be assigned to different experimental conditions. All AI-generated content is clearly marked and provided only as reference; final decisions are made by Area Chairs and the Program Committee.
Read the formal noticeComparison
| Aspect | Study 1: AI-assisted Peer Review Study | Study 2: AI Review Fairness Study |
|---|---|---|
| Main purpose | Study how AI tools can support reviewers in paper understanding, evidence location, note organization, and review preparation. | Study how AI-generated reference comments, scores, novelty analyses, and display conditions may affect human review behavior. |
| AI role | Provides substantial workflow support to reviewers throughout paper reading, including question answering, evidence location, note organization, and draft comment structuring. | Provides AI-generated reference information under controlled experimental conditions to study how such information may affect human review behavior. |
| AI content | Paper-based question answering, evidence location, review-note organization, and draft review support. | AI-generated review comments, reference scores, novelty analysis, and related reference information. |
| Relation to reviewer judgment | Supports reviewers in forming and expressing their own judgments, but does not independently assign scores, recommend acceptance or rejection, or define the final assessment. | Studies whether AI-generated comments, reference scores, novelty analyses, and different display conditions may influence reviewers' scores, rationales, novelty judgments, and recommendations. |
Participant Protection
Reviewer Participation and Protection
Reviewers will be informed of the relevant study conditions before participation. Reviewer-side research data, such as system usage records, interaction logs, or responses to study questionnaires, will be analyzed only in anonymized and aggregate form. Reviewer identities will not be publicly released, and participation or non-participation will not affect reviewers' standing in the conference review process.
Author Choice and Protection
Authors are encouraged to read the formal notices carefully before deciding whether to opt out of either study. NLPCC respects authors' choices. For both studies, research data will be analyzed only in anonymized and aggregate form. NLPCC will not publicly release authors' paper manuscripts, original review comments, author identities, reviewer identities, or any information that could be used to infer the identity of a specific paper, author, or reviewer.
Author Opt-out Instructions
Authors may opt out of one study, both studies, or neither study. Opting out of one study does not automatically opt the paper out of the other study. Authors who wish to opt out should send the corresponding email before the deadline.
Study 1 — AI-assisted Peer Review Study
support.reviewcopilot@gmail.comStudy 2 — AI Review Fairness Study
support.wismodel@gmail.comIn the email body, please include the paper ID, the name of the corresponding author, and the OpenReview link to the paper. Opting out of either study will not have any negative impact on the review, scoring, discussion, ranking, or final acceptance decision of the paper.
NLPCC 2026 Program Committee