24th European Semantic Web Conference (ESWC)

May 23-27, 2027 | Limassol, Cyprus

Policy on the use of Generative AI for Reviewers

Use of Generative AI: Reviewers

  1. Reviewers should not upload, submit, or otherwise share any part of a paper under review (including its title, abstract, text, figures, or data) to any generative AI tool or service, whether free or paid. Submitted papers are confidential documents, and their authors retain copyright over their unpublished work. Sharing them with third-party AI services risks breaching both author confidentiality and applicable copyright, regardless of the privacy terms or data retention policies of the service in question.
  2. Reviewers should not upload their review or part of it to any generative AI tools even for the sole purpose of rephrasing, correcting grammar, or improving the clarity of review text that they have themselves written.
  3. Reviewers may not use generative AI tools to perform or assist with the substantive analysis of a paper, to generate or summarise critiques or assessments of its contributions, or as a substitute for reading and properly evaluating the paper. Reviews must reflect the reviewer’s own independent, expert assessment.
  4. Reviewers are fully responsible for the accuracy, fairness, and content of their reviews, regardless of any editorial assistance obtained. Any errors, misrepresentations, or inaccuracies introduced through the use of generative AI tools remain the responsibility of the reviewer.
  5. Reviewers must not cite polished or fluent writing as evidence of improper generative AI use in a submitted paper. Formal or well-structured prose is not, in itself, grounds for raising a concern about undisclosed AI-generated content. If a reviewer believes a paper has made undisclosed use of generative AI, this must be based on concrete textual cues, such as prompt artefacts (e.g. phrases like “let me think about that” or “certainly, here is a summary”), inappropriate emoji, or other content clearly inconsistent with human-authored academic writing. Any such concern should be raised with the programme chairs rather than reflected in the review score.

Send feedback on this page to eswc2027@easychair.org .

This page is adapted from the ISWC 2026 Review Guidelines , which include the ISWC 2026 policy on the use of Generative AI by reviewers.