A sharp increase in citations of statistician Andrea Riebler’s work led her to investigate a growing market for ready-made research tools. She and colleagues found evidence that companies in China are selling software, artificial-intelligence writing assistance and instructions designed to help users produce research papers from public health data in less than two hours.
Riebler, from NTNU’s Department of Mathematical Sciences, said the tools enable the rapid production of simple, standardized analyses. Although the results are not necessarily incorrect, she questioned whether every paper generated from the available datasets is scientifically relevant or useful.
The investigation began after citations to a 2017 paper by Riebler and University of Zurich professor Leonhard Held rose from 70 in 2024 to more than 250 the following year. Their paper described software for calculating age-specific forecasts of disease incidence and mortality, with source code and usage instructions made freely available.
Many of the newly identified papers were written by Chinese researchers and used data from the Global Burden of Disease project. They addressed subjects including stroke risk, testicular cancer, falls among older people in China and colorectal cancer. Riebler said the papers often shared similar titles, structures and graphics, with some appearing to use the model implemented in her software.
Riebler, Held and researcher Peter Methys Degen traced the pattern to a commercial software package mentioned in one of more than 700 GBD papers they examined. Their findings are reported in the Journal of Clinical Epidemiology in a paper titled “Paper mill or paper mine? A tentative answer to the sharp increase in research papers based on the Global Burden of Disease database.”
The researchers identified a company in the Guangzhou region advertising tools that could produce research funding applications within minutes and publication-ready scientific papers in under two hours. They also found two other software packages sold by Chinese companies for similar purposes.
The researchers said that combining these tools with AI-based writing assistance could create a new model for mass-producing research. The concern extends beyond the Global Burden of Disease database to other large health datasets that are freely available.
A central problem is that the source code is usually not shared. In 99% of the papers examined by Riebler and her colleagues, the code was unavailable, making it difficult to verify or reproduce the analyses. The researchers said this undermines transparency and accountability, which are fundamental principles of scientific research.
The volume of submissions is also placing additional pressure on academic journals. Peter Methys Degen told The Verge that the peer-review system was already struggling with the number of papers and the limited supply of reviewers, warning that easier mass production could push it toward a breaking point.
The journal Accountability in Research reported a 60% increase in submissions in 2026. Editor David Resnik told The Verge that many of the papers, ironically focused on research misconduct, were probably generated with AI and that the surge could seriously strain peer review.
Riebler, who serves on the editorial boards of two journals, said her own inbox was filling with submissions. She noted that the question is not only whether a paper was written by a language model or a person, but also whether the growing number of submissions can be properly assessed.
Riebler and her colleagues are calling for stronger monitoring of online platforms that sell ready-made research tools. Riebler also argued that sharing source code should be mandatory.
Sources: Phys.org.