
Submit AI Slop: Lose a Year
ArXiv Drops the Hammer: 1-Year Ban for Researchers Submitting AI Slop
TL;DR: ArXiv will ban researchers for a year if they submit papers with obvious AI slop, including hallucinated citations, LLM meta-comments, or unchecked GenAI output passed off as rigorous science.
CS section chair Thomas Dietterich confirmed offenders face a 12-month ban, plus a requirement that future submissions clear peer review first. The trigger? "Incontrovertible evidence" like fake references or leftover prompts such as "here's a 200-word summary, want changes?"
Tools like Turnitin's *Responsible AI in Action* prove detection and legitimate AI use can coexist. The issue isn't AI, it's lazy science.
Hopefully ArXiv's stance raises the bar for every publisher that cares about its readers and content quality.