Imagine if every doctor and nurse in Australia could access a patient’s complete health record in seconds, right at the point of care, identify emerging risks earlier, and minimise the administrative tasks that continue to consume a significant portion of their working hours. It would free clinicians to focus on what they do best: caring for patients.
AI can automate or augment up to half the tasks across the system. But that depends on the quality, accessibility and trustworthiness of the data that underpins it. Today, much of that data remains fragmented, poorly governed, and locked in silos.
Without a trusted data foundation, AI simply magnifies those limitations. That is why so many healthcare initiatives stall at the pilot stage. The challenge is not a lack of AI capability but rather, a lack of connected, governed and usable data.
Finding the balance between accessibility and privacy
For AI to be effective in healthcare, it must first be trusted. That trust is defined by how data is handled, especially personal and sensitive data used in healthcare. The same information that enables better decisions also carries significant responsibility around privacy, consent and security. If data is mishandled, the impact is immediate, affecting confidence and adoption of new technology.
Australians were reminded of this in 2022 when the Medibank breach exposed sensitive health data of millions of customers. It remains one of the largest data breaches in Australian history and continues to shape how people think when sharing their data. In fact, many Australians remain reluctant to share their health data without clear safeguards and informed consent.
Yet the cost of that caution is equally real. Without access to complete and reliable data, AI cannot identify patterns early, support timely interventions, or help health systems learn from past outcomes. Delays persist, opportunities are missed and care remains limited, probably reactive rather than predictive.
Healthcare leaders therefore face two seemingly opposing pressures: enabling access to data that can improve care, while ensuring privacy, security and consent are protected at every step.
Build trust into the data foundation first
Resolving this tension requires a different approach to governance. Rather than treating privacy, security and consent as controls layered on top of the data, they must be embedded into the way data is managed, shared and used from the outset. Governance must be seen as an enabler rather than a blocker.
This means governance follows the data, wherever it resides. AI operates within clearly defined guardrails, with consent applied in context, and organisations retain full visibility into how data is accessed and used. These capabilities are not administrative overheads but prerequisites for trustworthy AI.
Trust and governance are not simply compliance requirements. They are the foundations that allow organisations to unlock value from AI in practice.
The experience of the Hong Kong Hospital Authority illustrates why a trusted data foundation came first. Responsible for more than nine million patient records across 43 public hospitals and 30 source systems, the organisation first focused on creating a unified data platform capable of bringing together disparate healthcare data at scale.
The impact extended well beyond technology. Real-time visibility into patient flow helped reduce A&E access blocks from around 12 to 3 per cent. That same foundation is now supporting AI-driven initiatives designed to assist diagnosis and treatment, helping clinicians access better information and identify potential abnormalities earlier.
The sequence clearly matters: the data foundation came first then the AI followed, because the data was ready.
Bring AI to the data, wherever it resides
There is another reality the healthcare sector must confront as AI becomes more central to care delivery. Data does not sit in one place, and it never will.
Healthcare environments have evolved over decades. Data is distributed across electronic medical records, pathology systems, administration platforms, private cloud environments, public cloud services and on-premises infrastructure. In NSW alone, patient information sits across multiple systems without current statewide connection.
Years of investment, regulatory requirements and operational needs have created highly distributed ecosystems, and that is unlikely to change.
This challenge extends well beyond healthcare. Cloudera’s 2026 Data Readiness Index found that nearly four in five organisations say their AI initiatives are constrained by limited access to data across environments. Even in APAC, where organisations report stronger progress, just 27 per cent report their data sources are fully integrated.
The answer here is not to force everything into a single environment. Moving large volumes of sensitive data adds cost and risk without solving the underlying problem.
A more effective approach is to bring AI, analytics and governance to the data wherever it resides. This allows organisations to create a more complete, near real-time view of the patient while preserving existing investments and maintaining control. It also enables innovation without compromising security, privacy or operational resilience.
Ultimately, the question is not how much AI can be deployed; it is whether the system can trust the data that underpins it. Fragmentation and governance are not side issues. They determine whether AI scales and whether it reaches clinicians and patients in a meaningful way.
Healthcare AI success starts with clean data, not models
The opportunity for Australian healthcare is clear. We already have the data, but a significant portion of it is fragmented, inconsistently governed and underutilised. What comes next is building the trusted foundation that allows it to be connected, controlled and confidently used by AI at scale.
The future will not be defined by how much data is collected. It will be defined by how well existing data is made usable. When that happens, AI can move beyond pilots and become what it was meant to be. A practical, trusted assistant that supports clinicians, improves outcomes and helps the system act earlier when it matters most.
Viewpoint articles are the author’s opinion, produced without payment or sponsorship.
Add Comment