A quiet change is taking place in Australian doctors' offices: increasingly, the notes from a consultation are being drafted not by the doctor but by an artificial-intelligence "scribe" listening in. As the tools spread, Australian regulators and medical organizations are urging caution, warning that safeguards for patient privacy and consent must keep up.
What an AI scribe does
An AI scribe records the conversation between a doctor and patient, transcribes it, and produces a draft clinical note that the doctor is expected to check and approve before it enters the medical record. Supporters say the tools cut the time doctors spend typing and let them focus on the patient in front of them. The Royal Australian College of General Practitioners (RACGP) acknowledges those benefits, including reduced administrative load, while cautioning that the technology must be used carefully.
Adoption is climbing fast
Use of the tools among Australian general practitioners has risen sharply from a low base, according to surveys by the RACGP and medical indemnity insurers, with a growing share of GPs either using an AI scribe or actively considering one. That rapid uptake, running ahead of settled rules, is a large part of why regulators have begun to speak up.
The privacy warnings
The Office of the Australian Information Commissioner (OAIC) has warned organizations about entering personal information, including sensitive health data, into commercially available AI products that may not adequately protect it. Because a scribe captures health information in real time, it sits at the sensitive end of what privacy law is meant to guard.
The Australian Health Practitioner Regulation Agency (AHPRA) has stressed that clinicians remain responsible for anything AI produces that affects a patient's record, and that they must obtain proper consent. Recording a consultation without a patient's agreement can also fall foul of surveillance laws in parts of Australia. Medical bodies advise that patients be told when a scribe is in use, be able to decline without affecting their care, and have their data stored securely, ideally on Australian servers.
Accuracy, too
Beyond privacy, experts point to reliability. Reviews of AI-generated notes have found omissions of clinically relevant details and occasional "hallucinations," where a system produces confident but incorrect information. That is why the guidance is consistent on one point: the doctor, not the software, is accountable for the final note, and every draft should be reviewed before it is trusted.
The result is a familiar tension in health technology. The tools are genuinely useful and increasingly popular, but the rules on consent, data handling and oversight are still catching up, and Australian authorities are signaling that they intend to close that gap rather than let adoption outrun it.
