Study Warn AI May “Flood” Literature With Synthetic Research

AI is enabling the mass production of near-identical scientific papers that slip past plagiarism checks, raising concerns about the surge of  low-quality academic publications.

In a preprint posted on Nature, researchers revealed that 112 journals published more than 400 duplicate papers throughout the last 4.5 years.

These copycat studies, based on the US National Health and Nutrition Examination Survey (NHANES), reused the same data to report nearly identical findings.

“If left unaddressed, this AI-based approach can be applied to all sorts of open-access databases, generating far more papers than anyone can imagine,” says Csaba Szabó, a pharmacologist at the University of Fribourg in Switzerland.

“This could open up Pandora’s box [and] the literature may be flooded with synthetic papers,” Szabó added, as reported by Nature.

The researchers demonstrated how easy it is to misuse AI: they asked ChatGPT and Gemini to rewrite three redundant NHANES studies. “We were shocked that it worked straight away,” says co-author Matt Spick of the University of Surrey. The AI-written manuscripts evaded plagiarism detectors used by publishers.

“This shouldn’t be happening, and it doesn’t help the health of the scientific literature,” adds Spick.

Editors are worried that companies selling fake papers, also known as paper mills, could exploit this loophole. “These are completely new challenges for the editors and the publishers,” says Igor Rudan at the University of Edinburgh.

Publishers have started to implement new measures to address this issue. Frontiers flagged 32% of the identified papers saying these studies were published before their new integrity policies took effect. Springer Nature, whose journals carried 37% of the papers, vowed investigations.

“We take our responsibility towards maintaining the validity of the scientific record very seriously,” says Richard White, editorial director for Scientific Reports.

“AI-driven redundancy, in general, poses a serious and ongoing challenge to publishers,” says Elena Vicario, Frontiers’ head of research integrity.