Reviewers Found Fabricated Citations in 15 of 22 Submissions

Two machine-learning researchers published their summer review load: of 22 submissions across NeurIPS, WACV and an ECCV workshop, 15 contained fabricated citations, invented author lists for real papers, or writing that was unmistakably machine-generated. Two submissions that swapped real authors for imaginary ones were accepted for oral presentation, conditional on fixing the references. The surrounding literature agrees on the trend: an audit of 2.5 million biomedical papers found the share containing at least one fabricated reference rising from 1 in 2,828 in 2023 to 1 in 277 in early 2026. The authors released an automated bibliography-audit tool alongside the post.