Policy on LLM use for writing reviews or meta reviews

We permit limited and responsible use of AI tools to assist human writing of peer reviews and meta-reviews. 

However, it is mandatory to disclose use of AI tools. The reviewing form would solicit this information. If you use LLMs to generate or edit any portion of your (meta) reviews, you will be required to report your original, self-written paper assessment, and any LM interactions, in an accompanying textbox. (In doing so, consider skipping the LLM and submitting your original text directly! We, and the authors, are much more interested in your unedited thoughts than what an LLM has to say.)

Any LLM-generated falsehoods, hallucinations, or misrepresentations in reviews, or reviews that are otherwise low quality, are subject to disciplinary action as a failure to fulfill the reciprocal reviewing requirement. Any use of AI tools to populate the initial self-report, or substantial mismatches between the content of the original self-report and final reviews, will also be considered misrepresentations. We reserve the right to use both AI detection tools and original self-reports to identify errors resulting from inappropriate LM usage. Consequences may include desk-rejecting all papers authored by the (meta) reviewer or restrictions on submission to future conferences.

FAQs: 

Q: Will my initial unpolished reviews be shown to the authors? 

No. It may be used by the senior area chairs and program chairs to ensure the quality and integrity of the reviewing process. 

Q: What if I polish my reviews iteratively using a multi-turn conversation?

We request reviewers to share their input to the LLM across all turns.