AI Conferences and Academia: NeurIPS, ICML, ICLR — 2026-09-10
The AI academic calendar is intensifying this week with critical deadlines for ICLR 2027 and NeurIPS 2026 results approaching. A significant controversy has erupted over a data breach at OpenReview affecting ICLR 2026, exposing reviewer identities and revealing that approximately 21% of peer reviews were AI-generated. Meanwhile, South Korea is establishing new hubs for AI talent development ahead of hosting ICML 2026 in Seoul.
AI Conferences and Academia: NeurIPS, ICML, ICLR — 2026-09-10
OpenReview Breach Reveals AI-Generated Peer Reviews
A security flaw in the OpenReview API exposed the identities of reviewers for over 10,000 ICLR 2026 papers, representing about 45% of submissions. This breach facilitated bribery attempts and harassment of reviewers. Concurrently, analysis by Pangram Labs indicated that roughly 21% of ICLR 2026 peer reviews were fully AI-generated, raising serious concerns about the integrity of the review process.
Tight Timeline for ICLR 2027 and NeurIPS 2026
Researchers face a complex decision-making window this week due to conflicting submission policies. The regular paper deadline for ICLR 2027 is September 16, 2026 (AoE), while NeurIPS 2026 results are scheduled for release on September 24, 2026 (AoE). Because both conferences prohibit dual submissions of the same research, authors must decide whether to withdraw from NeurIPS before the ICLR deadline without knowing their NeurIPS outcome, or wait for NeurIPS results and miss the ICLR 2027 window.
South Korea Launches Elite AI Entrepreneurship Hubs
South Korea’s Ministry of Science and ICT has designated KAIST, POSTECH, UNIST, and Chosun University as key hubs for cultivating master’s and doctoral-level AI entrepreneurship talent. This initiative aims to bridge the gap between academic research and commercial application, leveraging the country's strong technical education infrastructure. The move comes as Seoul prepares to host ICML 2026, positioning the nation as a central player in the global AI conference circuit.
Local view
In South Korea, local media outlets like Dong-A Science are highlighting the government's strategic push to integrate academic AI research with startup ecosystems. The designation of leading universities like KAIST and POSTECH as entrepreneurship hubs reflects a national priority to retain top AI talent within the domestic ecosystem rather than losing them to overseas tech giants. This local perspective underscores the growing tension between pure academic research and the commercialization demands of the AI industry.
Context & numbers
- AI Review Prevalence: Approximately 21% of ICLR 2026 peer reviews were estimated to be fully AI-generated by Pangram Labs.
- Exposure Scale: The OpenReview API flaw exposed reviewer identities for 10,000+ papers, accounting for ~45% of ICLR 2026 submissions.
- Submission Growth: ICLR submissions have nearly tripled in three years, rising from 7,304 in 2024 to 19,814 in 2026.
- Key Dates:
- Sep 16, 2026: ICLR 2027 Regular Paper Deadline (AoE).
- Sep 24, 2026: NeurIPS 2026 Results Announcement (AoE).
On the radar
- NeurIPS 2026 Decisions: The community awaits the September 24 release of NeurIPS 2026 acceptance decisions, which will likely spark further debate on acceptance thresholds given the recent controversies surrounding ICML and ICLR review processes.
- ICLR 2027 Submissions: Authors are scrambling to finalize submissions for the September 16 deadline, with many facing the "prisoner's dilemma" of dual-submission policies.
- Policy Reforms: Following the OpenReview breach and AI review scandals, expect announcements from major conference organizers regarding stricter verification protocols and potential bans on undisclosed LLM usage in reviews.
This content was collected, curated, and summarized entirely by AI — including how and what to gather. It may contain inaccuracies. Crew does not guarantee the accuracy of any information presented here. Always verify facts on your own before acting on them. Crew assumes no legal liability for any consequences arising from reliance on this content.
Powered by
