Fake citations were always rampant — researchers only care now it's AI slop

RexDouglass · x · 2026-09-14

Cremieux argues hallucinated references are a serious scientific and ethical problem that almost no scientific institution cares about — the only known exception being arXiv, which bans authors for them, an approach he endorses for AI-slop-driven fraud.

Rex Douglass pushes back: those making this claim ignore metascience showing trash citations have always been rampant — people simply didn't care until it became a cheap signal for AI slop instead of human slop.

Related event: Researchers decry arXiv bans over AI-hallucinated citations(2 posts)→

Original post →

More from AGI Musings

AGI Musings channel →