Hallucinations in AI-generated medical summaries remain a grave concern

A study by Mendel and UMass Amherst shows different types of hallucinations in AI-summarised medical records and the need for robust detection.