AI Hallucinations Threaten Academic Integrity with Fabricated References
AI-generated hallucinations are increasingly infiltrating academic work, posing significant challenges to the integrity of scientific literature. Maxim Topaz, an associate professor at Columbia University, discovered that AI tools had inserted fabricated references into his research paper. This incident led him to investigate the prevalence of such errors, revealing that over 4,000 fabricated references were found in nearly 3,000 biomedical papers. The issue has escalated since AI tools became more widely used in research, with the rate of fabricated references growing more than 12-fold over the past three years. These AI errors, known as hallucinations, occur when AI models prioritize word patterns over factual accuracy, potentially undermining the scientific process by compromising the evidence chain used in medical guidelines and clinical decisions. 
