AMA calls for stronger rules as AI use in healthcare grows
21 Oct 2026
Speaking to the Joint Select Committee on Artificial Intelligence on 21 October 2026, AMA President-elect Dr Michael Bonning said existing laws do not provide enough protection for the use of AI in medical settings.
Dr Bonning drew on his previous experience helping develop AI tools in the United States. He explained that many systems are designed for specific tasks rather than wider clinical use.
Doctors already use AI tools to record consultations and support clinical decisions. While these systems may improve efficiency, reduce pressure on staff and help patients access care, their answers can still be wrong for an individual patient. Dr Bonning said doctors must remain responsible for clinical oversight.
The AMA outlined five priorities for reform:
- Classify healthcare AI as high risk by default, with oversight based on the system’s independence and potential harm.
- Establish a dedicated health AI governance body led by clinicians.
- Create a clear legal framework that works with existing regulators, including the Therapeutic Goods Administration (TGA) and the Australian Health Practitioner Regulation Agency (Ahpra).
- Ensure liability reflects who controls the AI system.
- Protect patient information, with data stored in Australia where possible.
Dr Bonning also warned that the difference between administrative and clinical AI tools may not always remain clear. Software updates could change how a product is used, allowing a system originally designed for administrative tasks to influence clinical decisions.
He said regulation should focus on what a tool actually does, rather than how its developer describes it. Even an automated appointment system could create risks if it cancelled a visit needed for important patient monitoring.
The proposed governance body would work alongside existing organisations, including the AI Safety Institute, the Office of AI and the TGA, rather than duplicating their roles. It would help healthcare professionals understand how AI systems operate and how they may affect care in real clinical settings.
On responsibility, Dr Bonning stressed that AI should support rather than replace doctors’ judgement. He also highlighted the need to avoid judging clinical decisions unfairly with the benefit of hindsight and to involve patients in decisions when a doctor’s assessment differs from an AI recommendation.
Dr Bonning last gave evidence on AI in 2024. He told the committee that the need for action had become more pressing since then.
AMA calls for stronger rules as AI use in healthcare grows
Source: Australian Medical Association (AMA).