AITC 2026 — Reviewer Guidelines

Guiding Principles

A good review creates value for two groups:

  • Authors — by giving them actionable, constructive feedback to improve their work.
  • The community — by helping maintain a high standard of published research and fostering a healthy research culture.

Reviewers are expected to be polite, respectful, and professional throughout the entire review process. Poor reviews harm not just individual authors, but the field as a whole.


Best Practices

  • Be thoughtful. The submission may be from a first-time submitter. Your words carry weight.
  • Be useful. Write reviews that are helpful to authors, fellow reviewers, and area chairs alike.
  • Be specific. Vague criticism is hard to act on. Back up every concern with concrete reasoning.
  • Be timely. Respect deadlines. If you can’t complete a review on time, notify your area chair immediately.
  • Avoid discriminatory language or anything unrelated to scientific content and clarity.

How to Write a Good Review

  1. Read the submission carefully — take notes, highlight key contributions, and flag sections that need a closer look.
  2. Evaluate the claims — check that experiments actually test the stated hypotheses, verify proofs where relevant, and assess whether conclusions follow from the results.
  3. Consider related work — does the submission position itself appropriately in the literature? Missing citations are fine if they wouldn’t change the submission’s conclusions; flag them if they would.
  4. Write constructive feedback — every weakness you raise should be substantiated and, where possible, paired with a suggestion for improvement.

Review Form

Summary

Briefly describe the submission and its contributions in your own words. This is not a critique — the authors should broadly agree with your summary. Do not simply paste the abstract.


Strengths and Weaknesses

Provide a thorough assessment across four dimensions:

  • Soundness — Are the claims well-supported by theory or experiments? Are methods appropriate and results honestly reported? (Note: soundness and impact are separate — a technically sound submission can have modest contributions, and a high-impact idea still needs rigorous support.)
  • Presentation — Is the submission clearly written and well-structured? Does it adequately discuss related work?
  • Significance — Does the submission address an important problem? Could it influence future research or practice?
  • Originality — Does it offer new insights, methods, perspectives, or novel combinations of existing ideas? Originality doesn’t require an entirely new method — new understanding or evaluation of existing approaches counts too.

Ratings

Rate each dimension on a 1–4 scale. If you give a 2 (fair) or 1 (poor), your Strengths and Weaknesses section must clearly justify why.

Score Meaning
4 Excellent
3 Good
2 Fair
1 Poor

Dimensions: Soundness · Presentation · Significance · Originality


Limitations

Have the authors adequately discussed limitations and potential negative societal impact? If yes, simply write “Yes.” If not, offer constructive suggestions. Authors should not be penalized for being transparent about limitations.


Overall Recommendation

Score Verdict
6 Strong Accept — Technically flawless, exceptional impact, no ethical concerns
5 Accept — Technically solid, high impact in at least one area, strong evaluation
4 Weak Accept — Solid contribution others can build on, but with some limitations (use sparingly)
3 Weak Reject — Clear merits, but weaknesses outweigh them; revisions needed (use sparingly)
2 Reject — Technical flaws, weak evaluation, poor reproducibility, or unclear claims
1 Strong Reject — Well-known results, unaddressed ethical issues, or incomprehensible contribution

Reviewer Confidence

Score Meaning
5 Absolutely certain — very familiar with related work, carefully checked details
4 Confident, but not certain — minor gaps possible
3 Fairly confident — some parts may have been unclear or unchecked
2 Uncertain — likely missed central parts or relevant prior work
1 Educated guess — outside your area or very difficult to understand

Ethical Concerns

If you believe the submission raises ethical issues, inform an area chair. Relevant areas include:

  • Discrimination, bias, or fairness
  • Harmful potential applications
  • Privacy and security
  • Legal compliance
  • Research integrity (e.g., plagiarism)
  • Other

If flagging, please describe your concerns in detail.


Policy Acknowledgements

By submitting your review, you confirm that:

  • You have complied with the AITC LLM use policy for reviewing.
  • You have abided by and will continue to abide by the AITC Code of Conduct.