40 Percent of Doctors Use AI Scribes, but a Report Warns They Can Hallucinate Patient Data

A smartphone displays the icons of some of the main artificial intelligence based apps, including Meta AI, Grok, Gemini, Perplexity, DeepSeek, and ChatGPT, in Saint-Mande, France, on July 15, 2026. Martin Lelievre/AFP via Getty ImagesArtificial intelligence “scribes” used in medical consultations, can introduce false information into patient records prompting incorrect diagnoses, a new report warns.The tool is often used to record conversations between doctor and patient, transcribing the exchange, and then creating clinical notes.“While these tools may offer administrative benefits for medical practices and practitioners, they are increasingly being used in highly sensitive settings, including by psychiatrists, psychologists, and other medical practitioners,” said Digital Rights Watch in its report, “AI Scribes in Australian Healthcare.”“The current regulatory environment in Australia is inadequate relative to the serious nature of the use case, and the systemic risks of the technology have neither been considered nor addressed.”In some cases, doctors have reported scribes omitting, misrepresenting, and even “hallucinating” clinical information.In one example outlined in the report, a scribe reported a patient’s breast cancer on the wrong side of her body.In another case, the word “not” was dropped from the phrase, “does not have epilepsy,” suggesting the patient had the condition when they didn’t.One doctor reported a scribe recording a neurological exam even when one wasn’t conducted.Another major issue is where patient data is being stored, and whether patient data could be used to train AI models.The report comes as AI scribe take-up soars among Australian GPs with 40 percent reporting using the tools in November 2025, double the volume in August 2024 (22 percent).Currently, AI scribes are not regulated if they only provide transcription services.A Therapeutic Goods Administration spokesperson told The Epoch Times that the technology would be regulated if it “has capabilities beyond transcription and summarisation.”“For example, analysing or interpreting clinical conversations by generating a diagnosis, differential diagnosis or treatment recommendation not explicitly stated by the healthcare practitioner,” the spokesperson said.“The Therapeutic Goods Administration (TGA) is stepping up its efforts to make sure that software-based medical devices meet Australia’s regulatory requirements,” the TGA wrote on Aug. 29, 2025 on its website.Meanwhile, the Royal Australian College of General Practitioners (RACGP) has issued its own statement on the risks of the technology.“An AI tool can provide unsafe advice because it has been poorly programmed, trained with inadequate or unrepresentative (biased) data, or is vulnerable to hacking,” the RACGP said in its statement.“There is also a risk an AI tool is used in an inappropriate setting or context, which can skew the device’s decision-making ability over time.”In the meantime, the college says doctors must obtain patient consent, update privacy policies, and investigate where patient data is stored.The Australian Medical Association (AMA) expressed similar views, noting that while AI can reduce inefficiency, it cannot replace clinical decision-making.“Regulation must ensure that clinical decisions are made with specified human intervention points during the decision-making process,” the AMA told The Epoch Times citing a previous position statement.The body also called for patients’ rights to be at the centre of AI development, and called for a national governance body involving politicians and the medical and legal profession.