New arXiv policy: 1-year ban for hallucinated references

· books · Source ↗

TLDR

  • arXiv’s Code of Conduct now holds authors fully responsible for all paper contents, including LLM-generated hallucinated references, with a 1-year submission ban as penalty.

Key Takeaways

  • Thomas Dietterich announced arXiv’s enforcement stance: signing a paper means full responsibility for contents regardless of how they were generated.
  • Penalty is a 1-year arXiv ban, after which subsequent submissions must first clear peer review at a reputable venue.
  • The policy applies to all authors listed on a submission, not just the primary submitter.
  • Whether the updated enforcement language is live on the arxiv.org policies page is unconfirmed as of announcement.

Hacker News Comment Review

  • Enforcement mechanics are the core open question: no public detail on how arXiv will detect hallucinated references at scale, whether via automated DOI verification or manual spot checks.
  • Commenters flagged ambiguity in the post-ban condition: “reputable peer-reviewed venue” is undefined, leaving room for venue-quality disputes.
  • The root tooling problem remains: BibTeX generation across arXiv, publisher sites, and conference pages is fragmented, and better citation tooling (e.g., Zotero, zbib.org) could reduce accidental errors versus willful negligence.

Notable Comments

  • @btown: Confirms penalty structure and notes the policy page may not yet reflect the new enforcement rules.
  • @ElenaDaibunny: Flags that enforcement at scale is unsolved – manual or automated DOI verification both have serious gaps.

Original | Discuss on HN