Doctors Are Using AI Scribes — Here’s What That Means for Your Privacy

You sit down in the exam room, and before your doctor speaks, they tap a screen. “I’m going to use an AI tool to take notes today — it records our conversation and generates a summary for your file.” You nod, but a quiet worry sets in. What happens to that recording? Who else can hear it?

This scenario is becoming more common. AI scribing tools — software that listens to patient visits, transcribes them, and automatically drafts clinical notes — are being adopted by clinics around the world. But in July 2026, the Australian government issued a formal warning to doctors about the privacy and safety risks these tools carry.

What happened

Australia’s health and privacy regulators jointly released an advisory cautioning medical practices about using AI scribing tools without adequate safeguards. The warning followed reports of patient data being stored on servers outside Australia, unclear consent processes, and inaccuracies in AI-generated summaries that could affect medical records.

While the advisory did not ban the tools, it made clear that doctors are responsible for ensuring any AI service they use complies with the country’s privacy laws — including the Privacy Act and the My Health Records Act. The message was blunt: if you use an AI scribe, you must know exactly where the data goes, how long it’s kept, and whether it is used to train the AI model.

Why it matters for patients

The risks fall into three categories.

Data storage and sharing. Many AI scribing tools are cloud-based. Your conversation might be processed on servers in another country, subject to different data protection laws. Even if the clinic promises encryption, the recording — containing your name, symptoms, medication history, and possibly mental health details — is now on a third party’s infrastructure.

Accuracy. AI transcription still makes errors. A missed “no” or a misinterpreted phrase could lead to an incorrect record. If that record is used for future treatment or insurance, mistakes may be hard to correct. The government warning noted that some doctors have discovered errors in AI-generated notes days later, by which time the original recording may already have been deleted.

Consent. In many cases, patients are asked to agree verbally or via a generic form. They may not be told whether they can decline, or what happens if they do. The advisory stressed that consent must be informed and voluntary — not a checkbox you barely glance at.

How to protect your medical data

You don’t need to refuse medical care, but you can take a few straightforward steps.

  • Ask before the appointment begins. “Will you be using an AI tool to record or transcribe this visit?” If yes, ask for the name of the tool and what company provides it.
  • Request the privacy policy. The clinic should be able to tell you where your data is stored, for how long, and whether it is used to train the AI. If they can’t answer, that’s a red flag.
  • Opt out. You have the right to say no. The doctor will still take notes — the old-fashioned way. If you feel pressured, consider switching practices.
  • Review your visit summary. After the appointment, ask for a copy of the AI-generated note. Read it carefully for errors. Most clinics have a correction process.

Keep in mind that the effectiveness of these steps depends on how transparent your doctor is. Unfortunately, not all practices are fully aware of the risks themselves. The Australian government warning is partly meant to push them to learn more.

The bigger picture

AI scribes can save time and reduce burnout for doctors. That’s a genuine benefit. But the technology is still maturing, and regulation is lagging behind adoption. Other countries are watching Australia’s move closely — similar warnings may follow elsewhere.

For now, the safest approach is to treat AI in the exam room like any other recording device: you should know it’s on, why it’s being used, and what happens to the tape.

Sources

  • Digital Trends, “Australian government warns doctors over AI scribing tools as privacy and safety concerns grow,” July 5, 2026.
  • Australian Government, Office of the Australian Information Commissioner, advisory on AI scribing tools (July 2026).