Particle.news

Download on the App Store

Paper-Mill Output Now Outpaces Retractions, Driving Publishers to AI Screening

Publishers have launched mass retractions with AI screening in an effort to catch fabricated papers before fraud networks overwhelm the scientific record.

Image
Image
Image

Overview

  • A PNAS study by Amaral and Richardson shows paper mills are doubling fraudulent publications every 18 months, while legitimate articles double only every 15 years.
  • Corrective measures such as retractions and journal de-indexing are rising at a much slower pace, doubling every 3.5 years and expected to address just 15–25% of fraudulent outputs.
  • Frontiers Media, Springer Nature and other major publishers have issued record-size retraction batches after uncovering organized networks that manipulated images, citations and sham peer review.
  • Indexing services including Web of Science and Scopus have deployed proprietary AI tools since 2023 to flag anomalous content and remove suspect journals from their databases.
  • Researchers and integrity experts are calling for enhanced editorial oversight, restructured publication incentives and stronger AI governance to safeguard the credibility of the scientific record.