arXiv is a curated research-sharing platform that promotes scientific innovation and equitable access to scholarship worldwide. Since its founding in 1991, arXiv has been freely available and open to all. arXiv now hosts over 2.9 million scholarly articles across eight subject areas.
As a pioneer in digital open access, arXiv.org is tightening its submission rules around AI-generated papers and fabricated citations is one of the clearest signs yet that generative AI is beginning to stress-test the foundations of scientific publishing. As language models become more capable at producing convincing academic writing, the line between legitimate research assistance and fully synthetic misinformation is getting harder to detect.
The bigger issue is not just fake citations — it is scale. AI can now generate enormous volumes of plausible technical content faster than human reviewers can verify it. That creates a new credibility problem for academia, especially in fast-moving fields like machine learning and biotech where preprint culture already moves at extreme speed. The next phase of scientific publishing may depend less on producing information and more on proving authenticity, sourcing, and human accountability.