Low-quality papers based on public health data are flooding the scientific literature
Briefly

An analysis identifies a dramatic increase in poor-quality papers generated from large open-access health databases. Researchers attribute the surge to potential exploitation by large language models and paper mills, with publication numbers rising from about 4,000 to 11,500 between 2021 and 2024. This influx includes repetitive and template-like titles, suggesting a lack of research originality and quality. The study serves as an early warning to peer reviewers and editors about vulnerabilities in the publishing system, as it becomes challenging to sift through potentially low-quality contributions in the scientific literature.
The analysis indicates a surge in poor-quality scholarly papers generated by exploiting large open-access health databases using large language models or paper mills.
Between 2021 and 2024, papers using data from major health databases increased from approximately 4,000 to 11,500, demonstrating an unexpected growth trend.
Read at Nature
[
|
]