AI BriefingNew arXiv Policy: One-Year Ban for Hallucinated References
Key point
arXiv will impose a one-year ban for unverified LLM outputs.
Details
arXiv holds that once someone is listed as an author, they bear responsibility for the entire paper regardless of how it was generated.
If a paper contains inappropriate language, plagiarism, bias, errors, or incorrect references produced by generative AI, and there is clear evidence that the author did not verify them, the paper is considered untrustworthy.
The sanctions are as follows.
- A one-year ban from arXiv
- Subsequent submissions must first be accepted at a reputable peer-reviewed venue
Examples of clear evidence include the following.
- Nonexistent references, i.e., hallucinated references
- Meta-comments or draft annotations left by an LLM
- Traces of model output such as prompts to fill in experimental numbers
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.
Related
Toss Doctor V2 Released: Maximizing QA Efficiency with LLM-based Automated Testing and Self-healing