Australia's federal health department has raised formal concerns about AI scribe technology in GP clinics and specialist practices, citing minimal oversight, inconsistent patient consent, and data-sovereignty risks that clinicians and suppliers may not fully understand.

The concerns surfaced in Senate estimates briefing documents released under freedom of information laws. In February 2026, the department noted that AI scribes "have little oversight," and that some suppliers market their products as sitting outside existing regulatory frameworks or as privacy-compliant without being transparent about data processing or storage. An April 2026 briefing from the department's AI Expert Advisory Group acknowledged that while AI scribes reduce clinician workload and burnout, they carry the same accuracy limitations as other large language models—with real consequences for patient safety, clinical accountability, and digital health infrastructure integrity.

One specific risk flagged in the June 2026 FOI release: some suppliers may not realize their platforms transmit patient data outside Australia, which triggers Privacy Act obligations and potentially the My Health Records Act—neither designed to handle ambient audio recording.

A Technology Scaling Faster Than Governance

Adoption is accelerating rapidly. An RACGP survey found GP use of AI scribes nearly doubled from 22% in August 2024 to 40% by November 2025. Companies report their tools have been used hundreds of millions of times globally in the 18 months prior to July 2026. That speed outpaces regulation: the regulatory classification of AI scribes remains genuinely unclear.

Under the Therapeutic Goods Act, AI scribes count as medical devices only if they make clinical recommendations. Documentation tools that simply record and transcribe often sit outside this framework. That gap matters significantly. It means there is no mandatory pre-market accuracy assessment, no required reporting of problems or errors, and no baseline standard for how data should be handled. Australia's health regulator is now considering whether safeguards are needed, but has not announced a specific plan.

Regulation is not entirely absent. AHPRA's telehealth guidance, updated in October 2025, specifically addresses AI scribes, advising practitioners to ensure patients understand what the tool does and to check whether it is appropriate for the consultation. The RACGP advises GPs to verify AI scribe output for accuracy and to seek documented consent at the start of each appointment. The Office of the Australian Information Commissioner (OAIC) has been active too—its digital health annual report addresses AI scribes, and its staff have met with scribe providers. In May 2026, the OAIC presented on AI scribes at Privacy Awareness Week. What is missing is unified, binding regulation that turns these advisory positions into enforceable standards.

Consent and the Self-Censorship Problem

How practices obtain patient consent varies widely. The department noted that informed consent requires patients to understand both the benefits and limitations of the specific tool—not just a blanket acknowledgment that recording is happening.

Getting consent wrong carries real clinical risks. A Melbourne psychiatrist, reported in May 2026, refused to accept new patients unwilling to consent to AI note-taking and identified two concerns: the possibility of data breach, and the likelihood that patients who distrust the technology will self-censor during appointments. The second risk is clinically serious in psychiatry, where therapeutic work depends on a patient's confidence in privacy. It is less acute but still relevant in general practice, where patients may withhold information about substance use, mental health, or sexual health if uncertain how their words are stored.

The Consumer Health Forum CEO noted that the FOI documents show consumers and government are asking the same questions about AI scribes—an alignment that tends to accelerate rather than delay regulatory action.

What Sits on the Horizon

The department's 2025 Safe and Responsible Artificial Intelligence in Health Care review characterized AI scribes as "very low" risk in some use cases—a framing that, against the February 2026 Senate estimates briefing, appears to be shifting. The Victorian Department of Health published its own guidance on ambient AI scribes in June 2026. The RACGP updated its AI scribes guidance as recently as 29 June 2026.

The next frontier is already emerging. Vision-enabled AI scribes capable of processing video as well as audio are in development—a capability that would expand both clinical utility and privacy risk significantly.

The federal department has not announced a consultation timeline or regulatory instrument. What it has done is place on the public record—through freedom of information release—an assessment that the current position cannot continue. That is typically how Australian health tech regulation begins.