Aged care in Australia is under pressure. Workforce shortages. Quality concerns highlighted by the Royal Commission. Growing demand as the population ages. Rising costs.

AI is often proposed as part of the solution. But most AI discussion focuses on acute healthcare, not aged care. The challenges and opportunities in aged care are different.

Here’s where I think AI can genuinely help—and where we should be cautious.

Applications That Make Sense Now

Falls prediction and prevention. Falls are a major cause of injury and hospital admission for older Australians. AI systems that analyse movement patterns, vital signs, and environmental factors to predict fall risk are showing promise.

Some residential facilities are implementing sensor-based monitoring with AI analysis. Early results suggest high-risk periods can be identified, allowing preventive intervention. This isn’t surveillance for its own sake—it’s targeted risk management.

Medication management. Polypharmacy (multiple medications) is common in aged care. Drug interactions, inappropriate dosing, and missed medications cause preventable harm.

AI medication review tools can flag high-risk combinations, suggest deprescribing candidates, and monitor for adverse effects. This supports pharmacists and GPs managing complex medication regimens.

Cognitive assessment support. Regular cognitive screening is important but time-consuming. AI tools that analyse speech patterns, interaction responses, and task performance can support initial screening, flagging residents who warrant fuller assessment.

Staffing and rostering. Aged care workforce challenges are partly about numbers, partly about deployment. AI optimisation of rosters—matching skills to acuity, predicting workload, balancing preferences—can improve both efficiency and staff satisfaction.

Clinical documentation. Care staff spend significant time on documentation. AI-assisted documentation (voice-to-text, structured data extraction, automated summaries) could free time for direct care.

Applications Where I’m Cautious

Social companion robots. There’s interest in AI-powered companion robots for residents with limited social interaction. I have reservations.

Human connection matters. While robots might provide some benefit, they risk becoming substitutes for human interaction rather than supplements. If companion robots enable further staffing reductions, the net effect on wellbeing might be negative.

I’m not opposed to research in this area, but deployment should be thoughtful about what we’re optimising for.

Automated care decisions. AI that recommends specific care interventions based on resident data is higher-risk than AI that supports human decision-making. Aged care decisions often involve complex trade-offs between safety, autonomy, quality of life, and resident preferences.

Human judgment should remain central to care decisions. AI that provides information is different from AI that prescribes action.

Extensive monitoring. Sensor-based monitoring can improve safety, but it can also feel like surveillance. Residential aged care is people’s homes. Privacy, dignity, and autonomy matter.

Monitoring should be proportionate to risk and implemented with genuine consent. The goal is safety, not control.

The Workforce Question

Aged care workforce shortages are severe. AI is sometimes discussed as a partial solution—robots filling care gaps, automation reducing workload.

I’m sceptical about AI significantly reducing workforce requirements. Most aged care work is relational and physical. Feeding, bathing, dressing, comforting—these require human presence.

Where AI can help the workforce:

• Reducing administrative burden, increasing time for direct care
• Supporting less experienced staff with decision support
• Optimising deployment of limited staff resources
• Enabling remote expert consultation for complex cases

What AI can’t do:

• Replace human presence and relationship
• Perform physical care tasks
• Provide emotional support and connection
• Make the difficult judgment calls that good care requires

If AI is used as justification for further staffing cuts, outcomes will worsen.

Implementation Challenges Specific to Aged Care

Aged care faces unique implementation challenges:

Infrastructure limitations. Many aged care facilities have limited IT infrastructure. Wi-Fi coverage, network capacity, and device availability may be insufficient for AI applications.

Workforce digital literacy. Care staff may have limited technology experience. Change management and training need attention.

Fragmented sector. Aged care includes large providers, small providers, residential care, and home care. Scaling AI across this fragmented sector is harder than within integrated health services.

Regulatory uncertainty. It’s not always clear how AI medical devices interact with aged care quality standards. Providers need confidence about compliance.

Funding constraints. Aged care operates on tight margins. AI investment competes with other priorities, and ROI needs to be clear.

Resident consent and capacity. Many aged care residents have cognitive impairment. Consent for AI involvement in care requires careful attention to capacity and substitute decision-making.

Principles for Ethical AI in Aged Care

Respect resident autonomy. AI should support resident choice, not override it. Monitoring and intervention should reflect what residents (or their representatives) want.

Prioritise quality of life. Aged care isn’t just about safety—it’s about living well. AI that improves safety at the cost of quality of life might not be net-positive.

Maintain human connection. AI should augment human care, not replace it. Any application that reduces human interaction should be viewed with suspicion.

Ensure equity. AI benefits should be accessible across provider types and locations, not just available to well-resourced urban facilities.

Involve residents and families. Design and implementation should include aged care consumers, not just providers and technologists.

Getting Started

For aged care providers interested in AI:

Start with clear problems. What are your biggest challenges? Falls? Medication errors? Documentation burden? Staffing? Start with problems, then evaluate whether AI helps.

Pilot carefully. Small pilots with willing staff and residents provide learning before broader deployment.

Measure outcomes that matter. Efficiency gains are fine, but what matters is resident wellbeing, staff satisfaction, and care quality. Measure these.

Engage staff. Staff resistance can derail AI initiatives. Engage them as partners in design and implementation.

Consider partnerships. Individual providers may lack capability for AI initiatives. Sector bodies, research organisations, or provider collaboratives can enable shared learning and resources.

Aged care needs innovation. AI offers genuine opportunities, but requires thoughtful implementation that keeps residents at the centre.

---

Dr. Rebecca Liu is a health informatics specialist and former Chief Clinical Information Officer. She advises healthcare organisations on clinical AI strategy and implementation.