Witnesses tell Congress that paper mills and rising retractions are contaminating the scientific record

At a House Science subcommittee hearing, Kate Travis of Retraction Watch told members that retractions have become far more common and that the project’s database documents more than 64,000 retractions overall and over 5,000 in 2025 alone. "Retractions mark unreliable papers generated by AI, articles created for a business transaction rather than genuine research, and...papers containing honest error and those with evidence of misconduct," she said.

Travis and other panelists described multiple mechanisms that let bad studies spread: paper mills that produce fabricated manuscripts for a fee, predatory journals that collect fees without adequate peer review, and citation networks that can legitimize fraudulent work. She noted that U.S. retractions account for a smaller share of retractions than some countries and called for improved transparency around institutional investigations and federal oversight offices such as the Office of Research Integrity (ORI) and NSF’s OIG, which witnesses said are under-resourced.

Committee members pressed witnesses on concrete protections. Witnesses recommended: expanding investigatory capacity and transparency of findings, considering a national database of misconduct investigations to inform hiring and grant decisions, and using existing enforcement tools (FTC, SEC, False Claims) more aggressively against exploitative actors. The subcommittee requested additional written materials and detailed provenance for several of the high‑volume paper‑mill estimates discussed at the hearing.