The entities enabling scientific fraud at scale are large, resilient, and growing rapidly

A large-scale empirical study published in PNAS reveals that organized entities enabling scientific fraud—commonly called 'paper mills'—are growing at an alarming rate, with suspected paper mill products doubling every 1.5 years compared to the overall scientific literature doubling every 15 years. Analyzing data from PLOS ONE, Hindawi journals, IEEE conferences, and multiple literature aggregators, researchers identified anomalous editorial patterns: a small fraction of editors (0.25%) handled 30.2% of all retracted articles at PLOS ONE. Networks of duplicated images across articles reveal coordinated, industrial-scale production of fraudulent science. Paper mills practice 'journal hopping' to evade deindexing, and only ~28.7% of suspected paper mill products have been retracted. The study argues current punitive mechanisms are insufficient and calls for systematic, well-resourced detection and accountability infrastructure separate from journals and institutions that have conflicts of interest.