An economist says fabricated citations from large language models in academic papers extend common practices of copying references without direct verification

— Commenters still call publishing hallucinated references unacceptable.

Original post

JB#671@PLINZOPMCMatthew Cole|@MATTBENCOLE

A lot of scholars don’t understand why publishing LLM hallucinated bullshit is unacceptable because it’s only a degree of kind away from the bullshit already normalized in their fields. This economist just straight up admits it.

4:29 PM · May 18, 2026 View on X

REPLYDR#103Delip Rao e/σ|@DELIPRAO

@PMinervini @VladChituc Citing non-existing bibtex entries, clearly hallucinated author names, does drop confidence the authors’ epistemic responsibility.

3:55 PM · May 19, 2026 · 46 Views