Australian Patient Finds False Drug Use Claim in AI-Generated Medical Record
Rebecca Green attended her first appointment with a urologist in Australia and agreed to have the consultation recorded and transcribed by an artificial intelligence (AI) system. Later, after undergoing kidney stone surgery in March, she discovered a serious and false claim in her medical record stating she used psychedelic mushrooms. The letter from the specialist to her family doctor suggested her minor psychedelic use might explain previous kidney area bleeding. Green, who has never used such substances, was shocked and worried about the potential impact on her workers' compensation claim and insurance.
Following her complaint, the urologist apologized and corrected the medical record, attributing the error to a likely mistake during AI transcription or dictation. The doctor also said she would review the clinic's AI usage but did not provide details on changes made. Green expressed concern that other patients might not detect such errors, emphasizing the need for human oversight in AI documentation.
AI medical scribes, which transcribe doctor-patient conversations into structured medical notes, are increasingly used worldwide to reduce administrative burdens and improve doctor-patient interaction. A 2025 Australian survey found 40% of family doctors regularly use AI scribes, up from 22% in 2024. Research in the US indicates AI scribes can reduce physician burnout and cognitive load, but risks include inaccuracies, misattributions, and fabricated information.
Privacy advocates highlight that patients often consent to AI recording with minimal awareness, sometimes only via waiting room signs or online forms, raising concerns about informed consent and data security. Regulatory frameworks remain unclear, especially when AI systems generate diagnostic or treatment suggestions beyond mere transcription.
In Israel, AI transcription tools are being tested but face challenges with Hebrew medical jargon and accuracy. Professor Ran Blitzer of Clalit Health Services stresses that despite AI benefits, doctors must verify all AI-generated records due to inevitable errors such as omissions, additions, or incorrect details. Clalit developed an evaluation framework called Optica to monitor AI tools’ safety, privacy, and accuracy over time. This case underscores the critical balance between AI efficiency and the indispensable role of human review in healthcare documentation.
