Last updated: 12 November 2026
This policy explains how IJCAI handles corrections, retractions and expressions of concern, allegations of research misconduct, complaints and appeals, and post-publication discussion for the IJCAI Proceedings. IJCAI follows the guidance of the Committee on Publication Ethics (COPE), including the COPE Retraction Guidelines.
1. Principles
The purpose of corrections and retractions is to correct the literature and ensure its integrity, not to punish the authors. The editor of each year’s IJCAI Proceedings is the Program Chair of that year’s conference. Corrections are approved by the IJCAI corrections team (section 2). The editor makes the final decision on expressions of concern and retractions. Where the editor has a conflict of interest or is not available, the IJCAI Publication Ethics Committee decides instead. IJCAI, as publisher, supports any investigation, can make recommendations, and can issue notices jointly with the editor.
2. Corrections
- Corrections are limited to factual errors in a published paper whose main findings remain reliable, for example a wrong value in a figure or a table, or an incorrect author name.
- Requests to update, extend or improve a paper, such as new results, additional experiments, rewording or added references, are not accepted as corrections.
- To request a correction to a published paper, authors write to corrections [at] ijcai.org with the paper’s DOI, a description of the error, and a corrected PDF. Each request is reviewed by the IJCAI corrections team at that address, and a correction is published only after the team has approved it.
- The corrected paper is published under the same DOI, with a dated correction notice on the paper page that links to the original version, which remains available.
3. Retractions
The editor may retract a paper when they no longer have confidence in its results and conclusions. In particular, retraction may be warranted when there is:
- clear evidence of major errors, irregularities in the data or images, or any form of misrepresentation (for example fraud, identity theft or fictitious authorship) that compromises the reliability of the findings;
- findings published elsewhere without proper attribution, permission or justification, or material or data used without proper authorisation;
- unethical research practices, compromised peer review, or undisclosed conflicts of interest that could bias the interpretation of the work or the reviewers’ recommendations.
Retraction notices link to the retracted paper, identify it clearly by title and authors, are published promptly, and are freely accessible to all readers. The retracted paper stays online, clearly marked “RETRACTED”, and its DOI keeps resolving. Batch retractions may be needed when there is evidence of systematic manipulation of the publication process, for example by paper mills.
4. Expressions of concern
If the editor is unsure whether a paper is reliable, because information is insufficient, delayed or unobtainable, retraction may not be appropriate, and the editor may publish an expression of concern instead. An expression of concern is also published while an investigation that will take time is under way.
5. Withdrawal before publication
Authors may withdraw a submission at any time before the camera-ready deadline. Accepted papers that are not presented in person at the conference, both orally and as a poster, are excluded from the proceedings, unless exceptional circumstances are approved by IJCAI in advance.
6. Allegations of research misconduct
- Allegations are handled according to COPE guidance.
- Anyone, whether reviewer, reader or author, may raise a concern, with evidence, with the IJCAI Publication Ethics Committee at ethics [at] ijcai.org. Every allegation of plagiarism that comes to IJCAI’s attention is investigated, whether or not it involves an LLM.
- The committee acknowledges each concern within 10 days. It assesses the evidence, gives the authors the chance to respond, and may contact the authors’ institutions. It aims to reach a decision within 45 days; complex cases may take longer, and an expression of concern may be published in the meantime.
- Outcomes may include no action, a correction, an expression of concern, retraction, notification of the authors’ institutions or employers and other relevant bodies, and exclusion from a limited number of future IJCAI conferences. Information may be shared with other conferences in the AI consortium.
- Reviewers and editors must report any signs of possible misconduct by authors, reviewers or editors that they notice.
7. Complaints
Complaints about IJCAI’s publication process, for example the handling of a submission, a review or a delay, go to the Publication Ethics Committee at ethics [at] ijcai.org. A complaint about a Program Chair or a committee member is handled by someone with no conflict of interest, such as the IJCAI President, the Secretary-Treasurer or the Executive Director.
8. Appeals
Acceptance decisions of the Program Chair are final and cannot be appealed. Authors may appeal against a decision on a retraction or on research misconduct by writing to ethics [at] ijcai.org. Appeals are decided by the IJCAI Appeals Board, whose members were not involved in the original decision. The decision of the Appeals Board is final.
9. Post-publication discussion
IJCAI does not host comments on published papers. Readers who wish to discuss a paper should contact its authors directly. Concerns about errors or possible misconduct can be raised with the Publication Ethics Committee as described in section 6.
