The Rise of the AI Scribe
In the past couple of years, AI scribes have moved from a niche technology to a common tool in Australian clinics. These tools use large language models to listen to, transcribe, and summarise conversations between doctors and patients. Proponents argue
they are a powerful solution to clinician burnout, automating the time-consuming administrative burden of writing detailed medical notes. This allows doctors to focus more on the patient in front of them rather than a keyboard. According to a poll by the Royal Australian College of General Practitioners (RACGP), the use of AI scribes by doctors in Australia nearly doubled in just over a year, rising from 22% in August 2024 to 40% by November 2025. However, this rapid adoption has raised red flags for federal health authorities, who note the technology has had 'little oversight'.
The Regulatory Puzzle: Who Watches the Watchers?
The core of the issue is that AI scribes don't fit neatly into one regulatory box. Their function spans medical practice, data privacy, and consumer rights, creating a complex puzzle for authorities. Oversight is currently a patchwork handled by multiple bodies, including the Therapeutic Goods Administration (TGA), the Australian Health Practitioner Regulation Agency (Ahpra), and the Office of the Australian Information Commissioner (OAIC). The key question is whether an AI scribe is simply a transcription tool or a medical device. The answer determines the level of regulatory scrutiny. Currently, a scribe is only classified as a medical device—and therefore regulated by the TGA—if it serves a 'therapeutic purpose', such as suggesting a diagnosis or treatment plan. Many current tools are marketed as general-purpose scribes to avoid this classification.
The TGA's Role: A Focus on Clinical Safety
The Therapeutic Goods Administration's primary concern is patient safety. Its regulations are 'technology agnostic', meaning they focus on the intended purpose of a product, not the technology it uses. If a software product, including one using AI, is intended to diagnose, monitor, or treat a condition, it's considered a medical device and must be included on the Australian Register of Therapeutic Goods (ARTG) before it can be legally supplied. The TGA is actively reviewing digital scribes to determine if their features, especially any that analyse or interpret information, push them into the medical device category. Concerns have been raised that some scribes with features suggesting diagnoses or treatments might already be operating in breach of the Therapeutic Goods Act. A report on the TGA's review is expected in the coming months.
Beyond the TGA: Privacy, Practice, and Data Security
While the TGA assesses clinical risk, other bodies manage different aspects. The Office of the Australian Information Commissioner (OAIC) oversees the handling of patient data under the Privacy Act. This is a major concern, as health officials have noted some AI scribe suppliers may be unaware that their platforms send patient data to cloud servers outside Australia. This raises significant risks for data security. Furthermore, the Australian Health Practitioner Regulation Agency (Ahpra) holds doctors accountable. Regardless of TGA approval for any tool, the individual practitioner remains professionally responsible for the accuracy of medical records and for obtaining informed consent from patients before using an AI scribe. Some patients have reported feeling pressured to consent or finding another doctor.
A Split Approach: Balancing Innovation and Risk
This split oversight model allows different experts to focus on what they know best: the TGA on clinical efficacy, the OAIC on data privacy, and Ahpra on professional standards. The goal is to create guardrails without stifling innovation. However, critics worry it creates gaps. The federal health department has noted the technology's rapid growth has outpaced regulatory control. There are also concerns about the accuracy of the large language models themselves, which can make errors that affect patient safety. As regulators work to catch up, the onus remains on healthcare providers to be cautious, fully vet the tools they use, and ensure patients understand and consent to their role in the consultation room.
