[D] Is ICLR not giving Spotlights this year?

The International Conference on Learning Representations (ICLR), one of the most prestigious AI research conferences, appears to have made a significant change to its 2025 paper acceptance structure. According to submissions visible on OpenReview, the conference has designated papers only as Oral presentations or Posters, with no visible Spotlight category—the traditional mid-tier designation between top Orals and standard Posters. This follows two major controversies: a security breach in March 2024 where over 4,000 confidential peer reviews were leaked from OpenReview, and growing concerns about AI tools being used to generate or heavily polish paper submissions. The conference has not issued any formal announcement about eliminating Spotlights, leaving the research community to speculate whether this is a permanent policy shift or a temporary measure.

Researchers on forums like Reddit hypothesize the change could serve multiple purposes: simplifying the review process after the leak compromised trust, avoiding the difficult percentile calculations required for Spotlights amidst potentially compromised reviews, or creating a cleaner binary distinction between accepted papers as AI submission volume grows. If permanent, this structural shift would mark one of the most significant changes to ICLR's paper categorization in years, potentially increasing competition for Oral slots and reducing the prestige gradient among accepted works. The conference committee is expected to clarify its position as the review cycle progresses, with implications for how other top-tier venues like NeurIPS and ICML might handle similar integrity challenges.