arXiv Implements Stricter Peer Review Rules for AI-Generated CS Papers
arXiv Tightens Submission Rules Amid AI Paper Flood
The preprint platform arXiv has introduced significant changes to its submission requirements for Computer Science papers, specifically targeting review and position papers assisted by AI tools. Effective immediately, these paper types must complete formal peer review before being accepted for publication on the platform.
New Submission Requirements
Under the updated rules:
- Review Papers and Position Papers submitted to arXiv's CS category must provide proof of successful peer review
- Authors must attach journal references or DOI metadata demonstrating acceptance
- Submissions lacking this documentation will face automatic rejection
The policy specifically targets two paper types that have seen explosive growth due to generative AI tools capable of rapidly producing literature surveys and opinion pieces.
AI Acceleration Forces Platform Changes
arXiv administrators noted the platform receives hundreds of review papers monthly, many being little more than "annotated reference lists" without substantive research contributions. While submission volumes grow across all categories, the CS section faces particular challenges from easily generated AI content.
"The development of AI has exacerbated our existing paper overflow problem," stated an arXiv spokesperson. "We're seeing unprecedented volumes of machine-assisted submissions that don't advance research discourse."
Policy Clarification: Existing Standards Enforced
The platform emphasizes this isn't a policy change but rather stricter enforcement of longstanding guidelines. Historically, arXiv accepted exceptional review/position papers through manual reviewer discretion when they demonstrated:
- Exceptional quality
- Broad community interest
- Authoritative authorship (senior researchers/scientific societies)
With AI lowering barriers to paper production, arXiv now requires external peer validation as a quality filter.
Implementation Goals
The new rules aim to:
- Elevate academic standards: Ensure readers access expert-vetted content
- Optimize resources: Reduce volunteer reviewer workload on marginal submissions
- Accelerate dissemination: Prioritize processing for novel research findings
The policy exempts papers analyzing technology's societal impact (cs.CY/physics.soc-ph categories). arXiv remains open to extending similar requirements to other categories if needed.
Key Points:
- arXiv mandates pre-publication peer review for CS review/position papers
- Change responds to surge in low-quality AI-generated submissions
- Platform seeks to maintain quality while managing growing volumes
- Exceptions remain for select paper types analyzing societal impacts