As the Australian healthcare sector continues to work out how to use AI, St Vincent’s has set up a council to guide the safe and ethical use of artificial intelligence. 

St Vincent’s, one of Australia’s largest health and aged care organisations which operates two inner-city health networks in Melbourne and Sydney, has launched a national AI Clinical Council to help guide the safe, ethical and effective use of artificial intelligence across its health and aged care services.

“Australia has a rare opportunity to become a global leader in responsible healthcare AI. We have world-class clinicians and researchers, internationally significant health datasets and high levels of public trust in our healthcare institutions. The question is whether we will use those strengths to build local capability or become increasingly reliant on technologies developed overseas,” said chief executive Chris Blake in a post on LinkedIn. 

The national Council will advise St Vincent’s on how to identify, evaluate and scale AI technologies that improve patient care and clinician wellbeing aligned with the organisation’s mission, values and commitment to compassionate care. 

“Artificial intelligence has the potential to transform healthcare, but only if it is implemented in ways that strengthen, rather than diminish, the human relationships at the heart of care,” said St Vincent’s chief medical officer, Chris Robinson. 

“The purpose of the St Vincent’s AI Clinical Council is to bring together clinical expertise, operational experience and community perspectives to help identify where AI can make the greatest difference for patients, residents and staff,” he added. 

Grappling with AI

St Vincent’s move comes as the Australian healthcare sector continues to grapple with AI. 

Last week, Australian Medical Association (AMA) president-elect Michael Bonning appeared before the Joint Select Committee on Artificial Intelligence, where he advised the committee that Australia’s current laws are not adequate for AI in healthcare.

He said doctors already use AI scribes and decision-support tools to inform diagnosis and treatment. He cautioned that even when applied safely, lifting productivity, easing workforce pressure, and improving care access and consistency, AI output can look technically accurate but still be wrong for the patient in front of the doctor, so clinical oversight must remain central in how it is governed. 

In addressing the inadequacy of the current AI laws and bringing about reform, Bonning outlined AMA’s priorities: treating healthcare as high risk by default, with oversight scaled to autonomy and potential harm of the AI tool; establishing a dedicated and clinician-led health AI governance body; legislating an overarching framework working through trusted regulators; ensuring liability follows control; and most paramount is protecting patient data, with Australian and, where feasible, onshore storage.

Bonning said that a dedicated health AI governance body would complement rather than duplicate the AI Safety Institute and Office of AI, and would continue to work alongside the Therapeutic Goods Administration to ensure clinicians are clear not only on how an AI tool works, but how the tool interacts within the clinical environment they operate in. This would ensure that healthcare professionals, technology developers and consumers can work in the same environments together to provide those multiple lenses into AI governance in health care.