ArXiv is implementing a one-year submission ban for researchers who submit low-quality, AI-generated content. The policy targets automated "slop" that degrades repository quality without advancing research.
Institutional gatekeeping around AI outputs creates friction in the submission pipeline. As language models lower the cost of generating plausible-sounding papers, venues face a scaling problem: manual review becomes the binding constraint. ArXiv's enforcement signals that quality filters—not just access controls—will become standard infrastructure. This shifts incentives away from volume-based submission strategies toward selective, higher-signal contributions.
For builders deploying AI-assisted research tools, the operational change is straightforward: output filtering and human validation become non-optional components of research workflows. Venues may adopt automated detection systems upstream, making submission infrastructure itself a point of leverage. Researchers using AI for drafting now face explicit reputational costs, potentially creating demand for tools that integrate quality assurance rather than pure generation speed. The ban creates a market differentiation layer between "generation at scale" and "generation with institutional compliance."