One of Australia’s largest not-for-profit health providers has thrown its weight behind artificial intelligence in the clinic, calling on governments to stop treating the technology as a series of one-off experiments and start building the policy, funding and trust frameworks that would let it scale safely across the health system.
According to IT Brief Australia, St Vincent’s Health Australia has made the case that clinical AI is already proving its worth in areas such as medical imaging, documentation and administrative load, and that the bigger risk now is inaction rather than overreach. The message, in essence, is that the country has run enough trials to know the technology works in narrow settings, and the harder question is how to move it into everyday care without losing the public’s confidence.
From pilots to the mainstream
Australian hospitals and GP clinics have spent the past few years dabbling in AI, mostly at the edges. Ambient scribes that draft consultation notes, algorithms that flag suspicious scans for a radiologist to review, and chatbots that handle routine patient queries have all moved from novelty to quiet routine in pockets of the system. What has been missing is any coherent national approach to how these tools are approved, funded, monitored and, crucially, trusted by the clinicians expected to use them.
That gap is the nub of the St Vincent’s argument. A health network running public and private hospitals, aged care and community services sees the same frustration play out repeatedly: a promising tool clears a local pilot, demonstrates a benefit, and then stalls because there is no clear pathway to reimbursement, no settled view on liability, and no shared standard for proving it is safe at scale. Each new deployment ends up relitigating the same questions, which is slow and expensive and tends to favour the biggest, best-resourced players rather than the best ideas.
The call lands at a moment when the broader health system is under acute strain. Workforce shortages, an ageing population and lengthening waiting lists have made the productivity promise of AI genuinely tempting to administrators who need to do more with the staff they have. If an ambient scribe can hand a doctor back an hour a day that would otherwise go on paperwork, that is an hour returned to patients, and across a large network those hours add up quickly.
Enthusiasm meets a hard dose of caution
Not everyone in the sector is convinced the answer is to accelerate. Clinicians and patient advocates have warned for some time that healthcare is precisely the wrong field for a move-fast culture, because the cost of a confident but wrong AI output is measured in misdiagnoses rather than bad restaurant recommendations. Concerns about algorithmic bias, the quality and provenance of training data, and the tendency of some models to produce plausible fabrications are well rehearsed, and they carry real force when the stakes are a person’s health.
There is also the question of trust between doctor and machine. Surveys of Australian clinicians have repeatedly found that many are open to AI as a support tool but deeply wary of anything that feels like it is making decisions over their heads or eroding their professional judgement. Regulators have been feeling their way carefully too. The Therapeutic Goods Administration already treats a range of software as a medical device, and privacy regulators have been sharpening their focus on how sensitive health data is collected and used to train these systems.
Seen in that light, the St Vincent’s intervention is less a demand to rip up the guardrails than a plea to build better ones. The argument is that clear national rules and dedicated funding would actually make AI safer, because a proper framework forces transparency, ongoing evaluation and accountability in a way that ad hoc pilots never will. Left to drift, the sector risks the worst of both worlds: slow adoption of genuinely useful tools alongside unregulated creep of riskier ones.
Why this matters for Australia
For Australia, the stakes are both practical and strategic. The health system is one of the largest employers in the country and one of the biggest line items in every state and federal budget, so even modest productivity gains have national consequences. A coordinated approach to clinical AI could ease pressure on emergency departments and regional services that struggle to attract specialists, and it could help close the stubborn gap in access to care between the cities and the bush, where a well-validated diagnostic tool might extend scarce expertise to places that will never have a resident radiologist.
There is a sovereignty angle as well. Much of the most capable AI tooling is built offshore, and Australian health data is among the most sensitive information any citizen holds. A national framework would give the country a chance to set its own terms on where that data lives, how it is used and who is accountable when something goes wrong, rather than importing foreign products and foreign assumptions wholesale. It would also give home-grown digital health companies, of which Australia has a growing number, a clearer runway to build and sell into the local system before competing abroad.
The flip side is the cost of doing nothing. If Australia keeps treating healthcare AI as a permanent pilot program, it risks watching comparable countries pull ahead on outcomes and efficiency while local clinicians are left stitching together unsupported tools on their own initiative. That is the scenario St Vincent’s appears most keen to avoid.
What’s next
The practical test will be whether governments respond with anything concrete. The levers are reasonably well understood: reimbursement pathways through the Medicare and hospital funding systems, clearer device and data rules from the TGA and privacy regulators, national standards for validating and auditing clinical AI, and investment in the digital infrastructure and workforce training that any serious rollout would require. None of that is simple, and much of it sits across the awkward divide between Commonwealth and state responsibilities that bedevils so much of Australian health policy.
What the St Vincent’s call does is shift the conversation from whether healthcare AI belongs in the system to how the country intends to govern it. On the evidence so far, the technology is not waiting for permission. The open question is whether the policy settings will catch up fast enough to steer it, and whether governments will treat that as the national priority a major provider now says it should be.
Sources: IT Brief Australia.
