50,000 abstracts flood ICLR 2027, straining peer review
iclr 2027 received around 50,000 abstracts before deadline, far exceeding 19,500 from last year, straining peer review process.
the international conference on learning representations (iclr) 2027 has already received around 50,000 abstracts, with the deadline still a week away. in comparison, iclr 2026 saw about 19,500 valid submissions. the large number of submissions may be partly due to authors hedging their bets on neurips results and planning to withdraw if accepted there. the rapid growth is likely fueled by broader ai hype, increased corporate research spending tied to publication records, and the ease of generating papers with ai. a previous neurips analysis found that authors heavily used ai to write their submissions. iclr 2026 faced challenges with low-quality ai-generated submissions and reviews, which eroded trust in peer review. with an even larger number of papers expected this year, these issues may intensify.