Australia's health department has raised concerns about AI tools that automatically record and write up notes during doctor's visits. The concern: these tools operate with little oversight, patients are not being asked permission consistently, and there are questions about where patient data ends up.

These concerns came to light through documents released under freedom of information laws. The department noted in February 2026 that AI scribes have "little oversight," and that some companies market their products as not needing regulatory approval or as fully private—without being clear about how patient information is actually processed or stored. A follow-up assessment in April 2026 acknowledged that these tools do help doctors by reducing paperwork and stress, but they have the same accuracy problems as other artificial intelligence systems, which can affect patient care and the safety of Australia's medical records system.

One specific worry: some companies may not realize their AI tools send patient data outside Australia, which creates legal questions under Australian privacy laws that were written long before this technology existed.

Technology Growing Much Faster Than Rules

Doctors are adopting these tools fast. A survey found that GP use of AI scribes nearly doubled from 22% in August 2024 to 40% by November 2025. Companies say their tools have been used hundreds of millions of times globally over the past 18 months. But regulations have not kept up: nobody is quite sure whether these tools should be treated as medical devices that need approval before they are used.

Under Australian medical device law, AI scribes only count as medical devices if they actually make medical decisions. Tools that just record and transcribe often slip outside the rules. That gap means there is no requirement that someone test these tools for accuracy before they are released to doctors, no system for reporting problems when they go wrong, and no standard rules for how the tools should handle patient data. The health regulator is now thinking about whether it needs to create new rules, but nothing has been announced yet.

Some guidance already exists. Medical authorities have advised doctors to explain to patients what the AI tool does and to get their permission. They have also told doctors to double-check the notes the AI writes for errors. But there is no single, legally binding regulation that brings all this guidance together and makes it mandatory.

The Problem With Permission and Privacy

Different doctors ask permission in different ways. The department said that real permission requires patients to understand the benefits and drawbacks of the specific tool they are using—not just a quick "okay, we will record your visit."

Not getting permission right can harm patient care. A psychiatrist in Melbourne refused to use the AI tool for new patients who did not agree to it, and warned about two risks: patient data could be hacked or leaked, and patients who do not trust the AI might hold back information during their appointment. The second risk is particularly important in psychiatry, where patients need to feel safe sharing personal information. But it also matters in general practice—people may not talk honestly about drug use, mental health, or sexual health if they are unsure how their words will be stored or used.

Consumer groups note that patients and the government are asking the same questions about these tools, which usually means regulation will follow.

What Comes Next

The department's 2025 review initially described AI scribes as "very low" risk in certain situations—but the February 2026 briefing suggests that view is changing. State governments have started issuing their own guidance, and major medical organizations have updated their advice as recently as June 2026.

More complex versions of these tools are coming—AI systems that can process video and audio together, not just sound. That would make the tools more useful but also create bigger privacy risks.

The federal health department has not announced when or how it will create new rules. But it has made clear, through official documents, that the current situation cannot continue. In Australia, that is usually how new health technology regulations begin.