ArXiv will ban authors for a year if they let AI do all the work (techcrunch.com)

ArXiv, the prominent preprint research platform, has announced a new policy aimed at combating the careless use of large language models (LLMs) in scientific papers. Authors found to have submitted work without verifying the results generated by LLMs may face a one-year ban from the platform. The policy highlights a growing concern over the proliferation of low-quality, AI-generated content, with evidence such as "hallucinated references" triggering scrutiny. In recent months, arXiv has implemented additional measures, including requiring first-time submitters to have endorsements from established authors, reflecting its commitment to maintaining research integrity. This policy is significant for the AI/ML community as it emphasizes authors' accountability for their work, regardless of the tools used in its creation. By mandating that researchers ensure the accuracy and appropriateness of AI-generated content, arXiv aims to improve the overall quality of submissions and combat the rise of fabricated citations, particularly in fields like biomedical research. With arXiv's transition to an independent nonprofit, the initiative also positions the repository to better address emerging challenges in the research landscape, ensuring that responsible AI usage becomes a standard across scientific communities.