The ICLR 2027 AI conference has received approximately 50,000 abstracts, significantly exceeding the 19,500 submissions recorded for the 2026 event.

Key facts
- •ICLR 2027 has received approximately 50,000 abstracts before the upcoming deadline.
- •The 2026 ICLR conference received 19,500 valid submissions.
- •Corporate research compensation structures tied to publication records are contributing to the submission volume.
- •Previous issues at ICLR included the submission of AI-generated papers with fabricated citations.
- •Reviewers have increasingly utilized AI to handle the workload of evaluating large numbers of papers.
The International Conference on Learning Representations (ICLR) 2027 has already collected roughly 50,000 abstracts with the submission deadline still one week away. This figure represents a substantial increase over the 19,500 valid submissions received for the 2026 conference. Organizers expect the final count to decrease as some authors withdraw submissions following the results of the NeurIPS conference, though the total is still projected to remain significantly higher than previous years.
By the numbers
Factors Driving Submission Growth
The surge in submissions is attributed to broader AI industry hype and corporate research policies that link compensation to publication records. Additionally, the use of AI tools has enabled researchers to produce papers at a much faster rate. A previous analysis by NeurIPS indicated that authors are increasingly relying on AI to draft their submissions.
Impact on Peer Review Quality
The increase in volume raises concerns regarding the integrity of the peer review process. During the 2026 conference, organizers faced challenges with low-quality, AI-generated papers that contained fabricated citations. Reviewers also reported using AI tools to manage the high volume of submissions, a trend that is expected to continue or intensify given the current influx of abstracts.
Advertisement
This article was independently rewritten by ManyPress editorial AI from reporting originally published by The Decoder.


