I flagged two research papers for fake authors and both were accepted as orals
Points and comments are a snapshot, not live.
Two reviewers found 68% of 22 submissions had fabricated citations or LLM-generated content.
Caleb and Isaac reviewed 22 papers for NeurIPS, WACV, and TerraBytes, finding 15 (68%) contained fabricated citations, fake authors, or LLM-generated writing. Two papers with hallucinated authors were accepted as oral presentations, with instructions to fix the references. A Nature analysis found tens of thousands of 2025 publications likely contain invalid AI-generated references. Zhao et al. estimated 146,900 hallucinated citations in 2025, with 85.3% persisting in published versions. The reviewers propose desk-rejecting papers with fabricated references, and advocate for automated flagging tools.
What commenters are saying
Commenters largely agree that AI-generated slop is overwhelming peer review and devaluing academic publishing. Several note the irony of AI writing papers, AI reviewing them, and AI summarizing them, with humans increasingly automated out of the loop. One camp argues that the system was already broken before LLMs, and AI is just accelerating a necessary reckoning. Others suggest practical fixes: payment for reviewers, deposit fees for submissions, or reputation systems. A commenter points out that non-native English speakers benefit from AI writing assistance, while another observes that the real problem is semi-bogus claims that aren't false but feel wrong.