FAIR Summit is a professional research gathering. Everyone who takes part (authors, reviewers, speakers, organizers, sponsors, workshop leaders, attendees, volunteers and staff) agrees to this code at every session, dinner and online channel connected to the summit.

• Be respectful and considerate. Engage with ideas, not people.
• Give credit for other people's work, and don't share unpublished work or slides without permission.
• Respect off-the-record sessions and people who ask not to be photographed or recorded.
• Harassment, discrimination, intimidation and unwelcome attention are not tolerated in any form.

Research submissions: submit only your own work, keep your submission anonymous during double-blind review, don't contact reviewers or area chairs about a submission, and keep anything you review confidential. The dual-submission policy is in the Author Guidelines.

If something happens at the summit, speak to any organizer. Research integrity concerns go to the Program Chairs. Reports are handled promptly and confidentially, and FAIR Summit may take proportionate action, as set out in the Code of Ethics.

FAIR Summit is committed to rigorous, responsible, and useful research on model training, evaluation, human data, and related areas of AI. The Code sets out the ethical principles expected of authors, reviewers, speakers, organizers, sponsors, workshop leaders, and attendees.

• Pursue rigorous and honest research, and be transparent about how conclusions were reached.
• Treat human contributors responsibly and respect data provenance and permitted use.
• Consider foreseeable harms, evaluate both capability and failure, and avoid discriminatory practices.
• Respect intellectual contribution, privacy and confidentiality; disclose conflicts of interest.
• Use AI tools responsibly and maintain independence of research judgment.

Read the full Code of Ethics →

Review is double blind: reviewers do not see author names, and authors do not see reviewer names. Each paper is reviewed, then discussed between authors and reviewers and between reviewers and the area chair before final decisions. Reviewers look for:

• Relevance to model training and evaluation.
• Rigor: clear methods, honest results and stated limitations.
• Claims proportionate to the evidence. Negative results, replications and studies of failure are welcome.
• Usefulness: something practitioners can learn from or build on.
• Clarity of writing and presentation.

Reviewers disclose conflicts of interest and keep submissions confidential, including from AI tools (see the AI Use Policy). Detailed reviewer instructions are shared with reviewers before the review period.

Papers are submitted through the Submit a paper form. Talks go through Apply to speak, and you can put someone else forward through Nominate a speaker.

• Register your abstract by the abstract deadline, then submit the full paper by the paper deadline. Dates are on the Call for Papers page; all deadlines are 11:59 PM Anywhere on Earth (AoE).
• Include a title, an abstract and the workshop that fits best, plus keywords and a one-line TL;DR if you can.
• Keep the paper anonymous for double-blind review: no author names or identifying links in the paper itself.
• Link code or data where it exists, using anonymized links during review.

Page limits and formatting requirements are added to the Author Guidelines before the paper deadline. Subscribe for updates so you hear first.

AI tools may be used in research and research communication. Authors remain responsible for all work submitted under their names.

• Verify AI-assisted claims, citations, code, analysis and other substantive outputs.
• Disclose material use of AI in research methods, data generation, analysis or interpretation.
• AI systems cannot be listed as authors, and must not be used to evade authorship responsibility, fabricate evidence, generate false citations or obscure the provenance of research.
• Reviewers must keep submissions confidential and must not upload them to AI tools or services that retain or train on the content.

This policy follows the Code of Ethics (section 11). Questions go to the Program Chairs, listed on the Call for Papers page.

Submissions are anonymous for double-blind review: remove author names and identifying details from the paper.

Dual-submission policy: do not submit work that is under review at, or already published in, another archival venue. Preprints, including papers posted on arXiv, and presentations at non-archival workshops are permitted. While under review, avoid publicizing the paper as a FAIR Summit submission.

Accepted authors present in a talk session or in the poster hall, where papers, evals and datasets are on the table for discussion.

• Acceptance emails include your format, time slot and any camera-ready steps.
• Poster and slide specifications are shared with acceptances.
• Bring the artifact, not just the idea: a demo, a dataset or the eval itself makes the best conversation.

Questions go to the Program Chairs, listed on the Call for Papers page.