New research flags an AI-driven reliability threat in biomedical literature: fabricated references inserted into published or submitted work can propagate through citations and clinical guidance. A study described by Columbia University nursing associate professor Maxim Topaz and colleagues audited nearly 2.5 million biomedical papers and roughly 97 million PubMed Central citations. The team reported more than 4,000 fabricated references embedded across nearly 3,000 papers, with the prevalence of fake references rising sharply over time—suggesting a “vertical” increase after AI tools became more widely used in research workflows. In the first seven weeks of 2026, the researchers found one in 277 papers included at least one non-existent reference. For universities, journals, and research offices, the development raises immediate process questions for authorship verification, citation tooling, and editorial checks when AI assistance is involved in manuscript preparation.