We recognize that authors of academic works use a variety of tools in the research on which they report, and to prepare the report itself, from simple ones to very sophisticated ones. We note that tools may generate useful and helpful results, but also errors or misleading claims. Knowing which tools were used is relevant to evaluating and interpreting academic works.
In view of this, we:
At MEC 2027, we promote an AI-free review process. In short, any use of AI tools and LLMs in review drafting and polishing is prohibited. Any non-adherence to this policy might lead to disqualification from being a reviewer for future MEC conferences.
Compared to the AI usage policies for authors, the AI usage policy for reviewers is more strict to prevent any possible leakage of ideas or code of the submissions. Reviewers must keep everything relating to the review process confidential. Reviewers cannot share any materials of the submissions (e.g., paper, code, other supplementary materials) with LLMs. It is not allowed to use LLMs to fix grammatical issues or to smooth the writing of your reviews either. Language-related mistakes are acceptable and the content of the review is what counts, not its smooth delivery - it is better that we have typos and grammar mistakes but are 100% confident that the assigned reviewer wrote the review.
Reviewers are responsible for the quality and accuracy of their submitted review regardless of any tools, resources, or other help they used to construct the final review.
This statement is an adapted version of the arXiv policy for authors’ use of generative AI language tools and ISMIR's policy for reviewers. We reserve the right to amend this statement as discussions continue and evolve.