Researchers are raising concerns that artificial intelligence tools are generating fake or inaccurate citations in biomedical and medical research papers, potentially creating risks for patient care and scientific reliability. A recent audit of millions of biomedical studies found more than 4,000 citations referencing research papers that do not actually exist.
According to researchers, these fabricated references can become especially dangerous when they influence clinical guidelines used by doctors and healthcare professionals. Experts warn that AI-generated citations often appear highly convincing, making them difficult to detect during peer review or editorial checks. In some cases, AI systems reportedly attributed fake studies to real scientists or created entirely fictional research papers.
The issue reflects a broader problem with AI “hallucinations,” where chatbots confidently generate false or misleading information. Multiple recent studies have found that AI systems frequently provide inaccurate medical responses, fabricated evidence, or unsupported conclusions. Researchers say the problem is becoming more common as AI tools are increasingly used in academic writing, literature reviews, and scientific publishing workflows.
Experts are now calling for stronger verification systems, including automated citation checking and mandatory human oversight before AI-generated material is published in medical journals or government reports. Analysts warn that while AI can improve productivity in research, unchecked reliance on generative AI could undermine scientific trust, healthcare decision-making, and public confidence in medical literature.
