New Delhi: An audit of 2.5 million biomedical research papers has uncovered nearly 3,000 studies containing fabricated citations that do not exist in scientific databases, raising fresh concerns over academic publishing standards amid the growing use of artificial intelligence tools.

The papers, published between 1 January 2023 and 18 February 2026, were drawn from PubMed Central’s Open Access database, operated by the US National Institutes of Health.

The findings, published in a correspondence article in The Lancet, revealed that researchers identified 4,046 fake references across 2,810 papers from a pool of 97.1 million verified citations. The study found the rate of fabricated references had risen more than twelvefold since 2023, with the sharpest increase beginning in mid-2024 — a period that coincided with the rapid adoption of AI-powered writing tools.

Lead researcher Maxim Topaz, an associate professor at Columbia University’s School of Nursing and Data Science Institute, warned that the issue could have direct consequences for patient care.

“This discovery directly impacts patients as medical professionals make treatment decisions based on clinical guidelines,” Topaz said. “A medical professional or clinical guideline developer has no way of knowing that the evidence they are relying on does not exist.”

He added that one of the reviewed papers contained 18 fake references out of 30 citations, with some already being cited in other studies and systematic reviews used to inform clinical practice.

ALSO READ | Food addiction decoded: How chips, cola and snacks are designed to drive overconsumption

According to the researchers, the fabrication rate rose from around four fake references per 10,000 papers in 2023 to 51.3 per 10,000 papers by the final quarter of 2025, before climbing further to 56.9 per 10,000 papers in early 2026.

The authors urged publishers to strengthen reference-checking procedures during the submission process and called on indexing services to include metadata that would allow readers to assess citation accuracy.

They also recommended that research integrity databases create a dedicated category for fabricated references to improve monitoring and accountability.

The team further called on publishers to retrospectively screen existing papers and issue corrections or retractions where fake citations may have undermined research conclusions.