New Delhi: arXiv, one of the world’s largest repositories for scientific research papers, is introducing stricter penalties for researchers who upload
academic papers containing unchecked AI-generated content. The shift is part of a broader trend in the scientific field of the proliferation of low-quality research papers being written using generative AI tools, TechCrunch reports.
The preprint platform, which has been used by computer scientists, mathematicians and scientists from the field of physics, will now be more active in combating articles that clearly contain unverified content generated by chatbots. These can include AI-generated references, citations from memories, or remnants of AI prompts and chatbot instructions from the manuscripts.
Authors could face one-year bans
The platform will intervene when there is “incontrovertible evidence” that authors didn’t review the AI-produced material before publishing, according to Thomas Dietterich, a member of the computer science section of arXiv. The aim of the revised policy is to enhance the quality of research and the credibility of scientific publishing.
The new guidelines permit authors to be banned from arXiv for a year for violation. Once the suspension was over, only papers accepted by a well-known and peer-reviewed journal or conference would be allowed to be uploaded by those researchers.
AI tools are allowed, but responsibility stays with researchers
arXiv stated that it was not prohibiting the use of AI in research writing. During the drafting stage, researchers can continue to utilise chatbots or other AI tools. The site, however, emphasised that the authors are responsible for everything in their papers.
If AI-generated text includes plagiarism, inaccuracies, deceptive statements or biased information, there will be no escaping accountability for the researcher, Dietterich said. The report states that arXiv moderators will initially be notified of suspicious submissions, and section chairs will independently determine whether submissions violate policies before taking any action. Authors will have an appeal option with decisions as well.
Concerns grow over fake citations and poor-quality AI papers
The stricter policy comes in response to broader worries in the scientific community about the misuse of generative AI in research. Academic integrity is a growing concern, with several recent studies pointing to fabricated citations and misreferenced content in research papers, especially those in the biomedical field where AI writing tools are increasingly employed.
Scientists have also expressed concerns about nonsensical AI-written passages being inserted into published papers without their authors noticing when reviewing. In recent years, arXiv has implemented tougher quality control measures, such as a requirement for endorsement for new contributors.
arXiv prepares for structural transition
In addition to the policy changes, there is also a big organisational change taking place in arXiv. The platform has been run by Cornell University for over 20 years and is now making progress towards becoming an independent non-profit organisation.
The move is set to bring more flexibility and funding to arXiv, addressing the challenges brought by new developments in AI-generated research and ensuring scientific integrity in the era of generative AI.
