ICML Conference

To ensure compliance w peer-review policies, ICML has removed 795 reviews (1% of total) by reviewers who used LLMs when they explicitly agreed to not. Consequently, 497 papers (2% of all submissions) of these (reciprocal) reviewers have been desk rejected Details in blog post 👇

Yet to be checked: - dual submissions - thinly sliced contributions Thank you to all reviewers for your hard work as we proceed with the review process!

As reviews begin, #ICML2026 is down to 23843 submissions! 528 papers have been withdrawn, desk rejected, or otherwise removed for, e.g.,: - blank PDFs - length, formatting, anonymization issues - malicious prompt injection - misrepresentation of identity - claiming fake conflicts

@Kaszanas Reviews fully generated by LLMs would be a violation of the ICML peer review ethics policy. See Reporting and Enforcement on icml.cc/Conferences/20…

@icmlconf What will you do with the reviews clearly using LLMs to conduct a fully automated review?

There has been some online discussion of prompt watermarks in ICML submissions. tl;dr: - Yes, this is one of the *conference*'s (several) scientific integrity measures - Yes, it's not infalliable (but it still helps) - No, your paper won't be desk rejected as a result 1/4

@paul_cal x.com/icmlconf/statu…

ICML journal editors have added hidden prompt injections to every paper sent to reviewers, to detect when reviewers are using AI It secretly tells the AI to use 2 specific phrases in the review A reviewer found it & was about to desk reject, assuming the author did it

Guidelines on "Generative AI Considerations" have been clarified accordingly here: icml.cc/Conferences/20…. If you have further questions, don't hesitate to reach out to the program chairs at program-chairs@icml.cc. And happy Valentine's Day from ICML! ❤️ 4/4

We told ACs/SACs they shouldn't desk reject papers for this. While *all* prompt injection by authors was forbidden, due to potential ambiguities in the source, we won't be penalizing papers with LLM *detection* prompts (only those that try to affect LLM review decision). 3/4

#ICML2026 has released the review assignments. This may have been the most complex matching process ICML has executed to date, both in scale and in the number of constraints. The matching balances expertise, senior/junior reviewer composition for each paper, LLM reviewing policy compatibility, geographic diversity, and other fairness and load constraints---all while operating at unprecedented scale. Huge thanks to the organizing team and to our reviewers for stepping up. If you’re curious what large-scale constrained optimization looks like in the wild… well, you’re reviewing it. 😄

ICML 2026 Ethics Violation Reporting can be done via the following (anonymous) form: docs.google.com/forms/d/e/1FAI… You can also email the PC chairs directly (from your own email or an anonymous one) at this addres: program-chairs@icml.cc

Machine learning is an *international community*. Every community is based on mutual trust and respect. But communities are not robust to betrayals of these principles. Protect our community by reporting, e.g., review collusion or submission deanonymization via this form 👇

Deadline for proposing a tutorial at #ICML2026 is in four days (February 6). Don't miss out!

The people want to know: how many submissions made it over the finish line? 🏁 Answer: 🥁 ... ... ... 24371! This is the number of submitted PDFs. There will be more checks (& hopefully withdrawal of half baked submissions 👀), stay tuned to see how many enter review

This is a signal we use to identify papers needing more attention in review (alongside score variance and reviewer confidence). Being flagged for closer attention is largely neutral from an individual author's perspective, as it means more careful review, which most authors putting in genuine effort would welcome. That said, it is possible that it might have some effect depending on how many papers an author has, and we'll be analyzing the data collected from this year to measure the magnitude of any such effect.

Full guidelines here: #desk-rejection" target="_blank" rel="nofollow noopener">icml.cc/Conferences/20…

As you prepare your submission for #ICML2026, follow the rules to avoid desk rejection! This may happen if, e.g., you violate double blind or have hallucinated references ALL your submissions may be rejected if, e.g., you try to compromise the integrity of the review process