Skip to yearly menu bar Skip to main content


Area Chair (AC) Guidelines

This page provides guidelines for Area Chairs (ACs) for AISTATS 2027. It is adapted from the AISTATS 2026 AC guidelines and previous AISTATS practices.

Responsibilities of an AC

  1. Identify and flag potential desk-rejection cases (e.g., serious formatting or anonymity violations, dual submissions, missing AI Use Statements, or suspected hidden prompts; see the FAQ below).
  2. Help with reviewer assignments by recruiting expert reviewers and adjusting assignments when necessary.
  3. Monitor the review process and ensure that reviewers are responsive.
  4. Ensure that every submission receives at least three high-quality reviews.
  5. Write an initial meta-review before the author rebuttal period.
  6. Organize reviewer discussions after author feedback and revise the meta-review accordingly.
  7. Make acceptance or rejection recommendations based on the paper, human reviews, author feedback, discussions, the AI factual-correctness review, and your own assessment of the submission.
  8. Recommend accepted submissions for oral presentation where appropriate.
  9. Resolve issues during the review process by recruiting emergency reviewers, addressing concerns from reviewers or authors, and escalating unresolved cases to the Senior Area Chair (SAC) or Program Chairs (PCs).

Contact Info

Your assigned SAC is your first point of contact. If you encounter an issue that you and your SAC cannot resolve, please contact the Program Chairs, Aymeric Dieuleveut and Quentin Berthet, at aistats2027+pc@gmail.com. Questions about ethics or conflicts of interest should be escalated to the PCs.

If the issue is technical and related to OpenReview, contact the OpenReview support team at info@openreview.net.

Important Dates

Dates may be subject to change. All dates are Anywhere on Earth (AoE; UTC-12), end of day, unless specified otherwise.

  • Abstract submission deadline: September 29, 2026
  • Bidding phase: September 29, 2026–October 6, 2026
  • Full paper submission deadline (including all supplementary material): October 6, 2026
  • There is no separate supplementary material deadline this year.
  • Paper-reviewer initial assignments: October 12, 2026
  • Review period: October 15, 2026–November 5, 2026
  • Checking and soliciting emergency reviews: November 6–13, 2026
  • AI review released to ACs for the initial meta-review: November 8, 2026
  • Initial AC meta-review period: November 8–13, 2026
  • Reviews, AI review and initial meta-review released to authors; author rebuttal period: November 16–22, 2026
  • Author–reviewer discussion period: November 23–December 3, 2026
  • Reviewer–AC discussion period: December 4–approximately December 14, 2026
  • Final AC meta-review deadline: December 15, 2026
  • AC–SAC discussion period: December 16–approximately December 22, 2026
  • SAC initial decisions: approximately January 5, 2027
  • Paper final decision notifications: January 20, 2027
  • Journal-to-conference track submission due: to be announced
  • Conference dates: May 3–6, 2027

New in 2027

Initial meta-review: ACs will write an initial meta-review for every assigned paper between November 8 and November 13, before reviews are released to the authors.

AI review of submissions: every submission receives an AI-generated review from an LLM-based system operated by the conference. More details below.

Main Tasks

Below is a list of tasks, with detailed explanations, that an AC is expected to conduct.

  1. Registration on OpenReview
    • Make sure that your preferred email address is accurate and that your OpenReview profile is up to date, including your conflicts of interest, expertise topics, current affiliation, publication record, and DBLP link where applicable.
    • Import all of your papers into your OpenReview profile.
    • Check your spam folder regularly for messages from noreply@openreview.net and whitelist that address if necessary.
    • Read and agree to abide by the AISTATS Code of Conduct.
  2. Paper bidding
    • Papers are assigned to ACs and reviewers using a combination of bids and a matching algorithm.
    • Bid carefully and report any conflict of interest that was not detected automatically.
    • Be flexible about handling papers that are not an exact match for your expertise. Contact your SAC if an assignment is seriously inappropriate.
  3. Initial checks and possible desk rejection
    • Check assigned submissions promptly for serious formatting or anonymity violations, dual submission, a missing AI Use Statement, suspected hidden prompts, and other violations of the Call for Papers.
    • The main-text limit is 8 pages. References, the AI Use Statement, reproducibility checklist, and appendices do not count toward this limit. Exceeding the page limit results in desk rejection.
    • The AI Use Statement must appear immediately before the references, even if the authors did not use generative AI. Flag any submission without an AI Use Statement: desk rejection is not automatic, but the absence of the statement at submission time may be used as a reason to desk-reject.
    • Do not desk-reject a paper unilaterally. Flag suspected violations and consult your SAC and the PCs when needed.
  4. Reviewer assignment adjustments
    • Check that the reviewers assigned to each submission have appropriate expertise and no conflicts of interest.
    • Aim for four suitable reviewers with diverse perspectives so that every paper receives at least three complete, high-quality reviews.
    • Recruit expert or emergency reviewers early when assignments are weak, reviewers decline, or reviews are missing.
  5. Monitor the review process
    • Send reminders and follow up promptly on missing reviews.
    • Read every submitted review. Request a rewrite when a review is superficial, unsupported, disrespectful, or otherwise below standard.
    • Reviews evidently produced by an LLM violate the reviewer policy and count as missing reviews. Flag them and arrange a replacement review.
    • Before reviews are released, ensure that every paper has at least three high-quality human reviews and that inappropriate language has been removed.
  6. Initial meta-review
    • New in 2027: write an initial meta-review for every assigned paper between November 8 and November 13, before reviews are released to the authors.
    • The AI factual-correctness review will be made available to ACs for this preliminary meta-review. Human reviewers will not see it while preparing their initial reviews.
    • Summarize the state of the human reviews, identify agreements and disagreements, consider the AI review critically, and state clearly the key points that the authors should address in their rebuttal.
    • This assessment is preliminary. Do not make it sound like a final decision; it must be reconsidered after the rebuttal and discussions.
  7. Author–reviewer discussion
    • As soon as author responses are available, initiate a discussion for every submission and ask reviewers to engage with the authors’ answers.
    • Ensure that reviewers acknowledge the rebuttal, respond to substantive points, and explain any resulting score changes—or why their assessment remains unchanged.
    • During the discussion phase, reviewers and ACs can also see the AI factual-correctness review and the authors’ response to it. Treat these as inputs to the discussion, not as an independent recommendation.
  8. Reviewer–AC discussion
    • Lead the closed discussion among reviewers, summarize the main strengths and weaknesses, and work to resolve major disagreements.
    • Ask reviewers to distinguish issues resolved by the rebuttal from those that remain.
    • Follow up appropriately on confidential messages from reviewers or authors and escalate sensitive cases to your SAC.
  9. Final decision recommendation and meta-review
    • Revise the initial meta-review to account for the rebuttal, reviewer discussion, any updated reviews or scores, and the AI review and the authors’ response to it.
    • Submit the final meta-review and recommendation by December 15, 2026.
    • Base your recommendation on the paper and the substance of the reviews and discussion, not on a mechanical average of reviewer scores.
    • Explain the reasons for the recommendation clearly and provide constructive feedback, especially for rejected papers.
    • Contact your SAC early when a case is borderline, the reviews conflict, or your recommendation differs from the reviewers’ consensus.
  10. AC–SAC discussion
    • Be available after the final meta-review deadline to discuss borderline cases and help your SAC calibrate recommendations across papers.
    • Update the meta-review if the final decision changes so that it accurately explains the outcome to the authors.

Best Practices

  • Be responsive. Respect deadlines, reply promptly, and tell your SAC as soon as possible if you will be unavailable during a critical period.
  • Be proactive. Monitor review quality and discussion activity throughout the process rather than waiting until a deadline.
  • Be fair and independent. Evaluate the substance of each paper and do not let reviewer scores substitute for your own judgment.
  • Be professional and kind. Use respectful language and show empathy when reviewers face legitimate difficulties, while ensuring that the work is completed.
  • Respect conflicts of interest. Do not discuss a paper with anyone who may have a conflict, and report missed conflicts immediately.
  • Report unethical or suspicious behavior to your SAC and the PCs. The AISTATS 2027 confidential reporting form may also be used.

Writing the Meta-review

These recommendations are adapted from the NeurIPS 2025 AC guidelines (original guidance by Chris Williams and John Lafferty).
  • Focus on the content of the reviews rather than the numerical scores.
  • Assess the quality and relevance of each review rather than relying only on reviewer confidence scores.
  • Summarize the decisive strengths and weaknesses; do not merely repeat the reviews.
  • Explain whether and how the author response addressed the reviewers’ concerns.
  • Represent disagreements fairly and explain how you resolved them.
  • Disclose if information beyond the paper, reviews, author response, discussion, and official AI review influenced your assessment.
  • Give a detailed justification when overruling a unanimous reviewer recommendation.
  • Take a clear position on borderline papers.
  • Counter biases against sound work in less fashionable topics.
  • Write the meta-review so that the authors can understand the reasoning behind the final decision.

More tips: this blog post.

Use of Generative AI

Large language models (LLMs) cannot be used to write reviews or meta-reviews. Reviewers and ACs must not delegate their judgment or critique of a submission to an LLM, generate reviews or meta-reviews with an LLM, or enter submission content into non-privacy-compliant tools. Violations may lead to sanctions, including desk rejection of the violator’s own submissions. Refer to the Ethical Conduct for Peer Review section of the Reviewer Guidelines.

Authors may use LLMs and other generative AI tools for research and writing, but every submission must include a mandatory AI Use Statement and the authors remain fully responsible for all content. LLMs cannot be authors. Submissions whose writing is unclear, unfocused, or padded may be rejected or, in serious cases, desk-rejected. Refer to the Use of Generative AI by Authors section of the Call for Papers.

AI Review of Submissions

New in 2027: every submission receives an AI-generated review from an LLM-based system operated by the conference. It targets factual correctness only, including incorrect or unsupported claims, errors or gaps in proofs, inconsistencies among text, tables, and figures, mismatches between the described and reported experimental setup, and misattributed related work. It does not assess significance or novelty, does not produce a score, and makes no recommendation.

Human reviewers do not see the AI review while writing their initial reviews. It is first made available to ACs for preparing their preliminary meta-reviews. It is then released to authors with the human reviews at the start of rebuttal and becomes visible to reviewers during the discussion phase, together with the authors’ response. Consider it critically as one input among others. All recommendations and decisions remain the responsibility of the ACs, SACs, and PCs and must be grounded primarily in human assessment and discussion.

Confidentiality

Keep everything related to the review process confidential. Do not use ideas, code, or results from submissions in your own work until they are publicly available. Do not discuss or distribute submissions without prior approval from the PCs. Code submitted for review cannot be redistributed. Invite any external or emergency reviewer through OpenReview rather than sharing submission materials through another channel.

Nomination

Reviewers and ACs can be nominated using this Google Form.

Frequently Asked Questions

  1. How do I contact my SAC or the Program Chairs?

    Use OpenReview’s contact functions to reach your assigned SAC. If the issue cannot be resolved there, contact the Program Chairs at aistats2027+pc@gmail.com.

  2. What if a paper does not have the reproducibility checklist?

    Flag any submission missing the reproducibility checklist to your SAC and the PCs. Desk rejection is not automatic, but the absence of the reproducibility checklist at submission time may be used as a reason to desk-reject. If the paper is accepted, remind the authors that the checklist must be included in the camera-ready version.

  3. What is the page limit?

    The initial submission is limited to 8 pages of main text. References, the AI Use Statement, the reproducibility checklist, and appendices are excluded from this limit. Camera-ready papers may use 9 pages of main text. Exceeding the page limit results in desk rejection.

  4. Where must the AI Use Statement appear?

    It must appear immediately before the references. Flag any submission without an AI Use Statement to your SAC and the PCs. Desk rejection is not automatic, but the absence of the AI Use Statement at submission time may be used as a reason to desk-reject.

  5. Which citation style is allowed?

    Either numeric or author–year citations may be used for the initial submission. Accepted papers must use author–year citations in the camera-ready version.

  6. What else may constitute a serious submission violation?

    Examples include breaking anonymity, dual submission, unauthorized changes to the author list after the abstract deadline, serious template or page-limit violations, hidden prompts intended to manipulate LLM review, and attempts to evade submission quotas or reciprocal-reviewing requirements. Flag questionable cases for your SAC and the PCs rather than resolving them alone.