AI in Aged Care: What the Regulator's AI Use Means

AI in aged care has no rulebook of its own yet, but the regulator has published how it governs its own use of AI. The Aged Care Quality and Safety Commission's AI Transparency Statement, last updated on 16 July 2026, sets out what the Commission uses AI for, who is accountable and where a person must make the decision. It does not bind providers. It does give every residential and home care service a clear benchmark for the questions a board, an auditor or a resident's family could reasonably ask.

Key Takeaways

The Commission says its AI tools are not approved to make decisions or take administrative action without human intervention, and that AI does not determine compliance outcomes without human review.
The statement describes the Commission's own AI use and is not a rule for providers, but existing duties on rights, privacy and quality still apply to care that AI supports.
The Commission relies on three simple controls: a register of AI uses, a named owner for every system and the power to suspend, restrict or stop a tool.
From 10 December 2026, an organisation covered by the Australian Privacy Principles must disclose certain computer-made or computer-assisted decisions in its privacy policy.
AI in aged care already covers scribes, fall and behaviour monitoring, translation and companion robots, and the benefits depend on human oversight, staff training and a way to override automated outcomes.
Note-taking, rostering, search and drafting tools count as AI when they write, predict, rank or recommend, even if nobody labelled them that way.

Where is AI used in aged care today?

Aged Care Research and Industry Innovation Australia (ARIIA) told the ABC in April 2026 that typical uses include digital scribes for note-taking and documentation, multilingual communication supports and tools that help with clinical assessments. An earlier ARIIA fact sheet describes AI paired with cameras or wearable devices to detect when an older person may fall, have higher-than-usual pain or show changes in behaviour and daily routine. Some larger providers have built their own platforms, while others partner with technology companies.

Ageing Australia chief executive Tom Symondson told the ABC that some providers use AI-enabled robots to help lonely older people, and said the technology should not take time away from human carers. Governa's overview of the types of AI used in aged care gives a plain-language breakdown, and its article on AI cameras in nursing homes covers the privacy side of monitoring.

What are the benefits and the risks of AI in aged care?

The benefit providers cite most is time. ARIIA's Liana Donleavy told the ABC that removing documentation tasks lets frontline staff return to care and spend more time with residents and clients. That only holds if the tool is low-risk, staff want to use it and someone checks its output.

The risks are well documented. The ABC reported concerns from the eSafety Commissioner's office about unregulated AI and its potential to change behaviour in some users. Ending Loneliness Together chair Michelle Lim asked who is behind the algorithms in AI companions and how a concern would be flagged if someone is in trouble. University of Sydney researchers who examined how 33 agetech companies market AI products found aged care framed as failing, older people framed as data and care workers framed as a liability, according to Hello Leaders' report on the study. The researchers urged leaders to scrutinise what is being sold, and said AI works at best as a supporting tool when staff have time, training and the ability to turn off notifications that distract from care.

Automated decision tools carry their own risk. The Commonwealth Ombudsman's October 2026 investigation, Automation and oversight in Aged Care, examined the algorithm-based Integrated Assessment Tool used for Support at Home. It raised concern that there was no way for assessors, delegates or internal reviewers to change an automated outcome they considered clinically inappropriate. The lesson for providers is practical: wherever a tool decides or ranks, a person must be able to override it, and the person affected must be able to ask for a review.

What is the government planning for AI in aged care?

The Aged Care Data and Digital Strategy 2024-2029 includes an AI framework. Its action plan includes assessing existing AI research, guidelines and safety controls, and a pilot for up to 20 health sector professionals to test AI in providing better information, the ABC reported. Hello Leaders adds that the National AI Plan, released in December 2025, commits to strengthening privacy, consumer protection and accountability frameworks. Both are direction-setting documents rather than a provider rulebook, which is why the Commission's own approach is a useful guide. Governa's AI readiness checklist and its look at the ethics of AI in geriatric care help a service test where it stands.

What does the Commission say about its own use of AI?

The statement describes three patterns of use. The first is analytics for insights, such as regulatory intelligence and risk analysis. The second is workplace productivity, where staff use approved tools such as Microsoft 365 Copilot Chat to summarise information, draft content and support research. The third is support for human decision making, where AI can help staff compare options or recommend wording while the staff member stays responsible for checking the output and for every final decision.

Two lines matter most for providers. The Commission says its AI tools are not approved to make decisions or take administrative action without human intervention. It also says AI-supported analytics inform its oversight of compliance, but AI does not determine compliance outcomes without human review.

The Commission's AI Policy was approved on 10 December 2025, and staff must complete internal training before they use approved tools. The statement adds that the public does not currently deal directly with Commission AI systems, and that the Commission will update the statement if that changes.

How does the Commission govern that use?

The statement lists eight responsible AI principles. Most translate directly into evidence a provider can hold. The table pairs six of them with proof a service could keep.

Clear accountability: every AI system has an accountable owner - Evidence a provider could hold: A named owner for each tool in your AI register
Risk and impact assessment before use - Evidence a provider could hold: A dated assessment completed before go-live
Human oversight in proportion to risk - Evidence a provider could hold: Written points where a person reviews or overrides the output
Transparency and contestability - Evidence a provider could hold: A plain statement of AI use for staff, residents and clients, and a way to question an outcome
Reliability and security - Evidence a provider could hold: Supplier security and data handling checked and recorded in the contract
Monitoring, review and continuous improvement - Evidence a provider could hold: A review date on every tool and a trigger for earlier review

Beyond the principles, the Commission keeps an internal AI use case register covering approved, proposed and exploratory uses. It also has escalation pathways for AI concerns or incidents, including the ability to suspend, restrict or stop AI use if controls are not working. The statement names two accountable officials and a Chief AI Officer.

A provider's version can be much smaller. The pattern is what counts: know what is in use, know who owns it and know how to switch it off. Governa's guides to keeping an AI tool register and to approving an AI use case show how to set both up.

Does the regulator's statement apply to providers?

No. It describes the Commission's own use. In its September 2026 analysis, law firm Russell Kennedy makes the point that the framework applies to the Commission as regulator, but shows that AI governance is no longer a theoretical discussion. The firm adds that the new aged care framework does not expressly regulate these technologies, and that this silence is not a green light. The Statement of Rights, the Code of Conduct and the Quality Standards apply whether care is supported by traditional processes or by AI-enabled systems.

In practice, an AI tool does not change who is responsible. If a note-taking tool puts an error into a care record, the provider owns the record. If a rostering tool leaves a client without a visit, the provider owns the missed service. This article is general information, not legal advice.

What changes for privacy on 10 December 2026?

From 10 December 2026, the Office of the Australian Information Commissioner says Australian Privacy Principle 1.7 requires an organisation covered by the principles to add information to its privacy policy when three conditions are met. The organisation has arranged for a computer program to make a decision, or to do something substantially and directly related to making it. The decision could reasonably be expected to significantly affect an individual's rights or interests. And personal information about that individual is used by the program.

When all three apply, the policy must describe the kinds of personal information used, the kinds of decisions made solely by the program and the kinds of decisions where the program does something substantially and directly related to making them. Ask your privacy lead whether your organisation is covered, then test each tool that ranks, scores or allocates against the three conditions. The OAIC's transparency resources include a fact sheet to help with that test, and Governa's earlier guide to the Privacy Act AI deadline looks at the same date from a provider's side.

Where is AI already running in your service?

Many services use AI without calling it that. Russell Kennedy lists meeting transcription and documentation tools, workforce scheduling and rostering systems, automated communication platforms, decision-support tools, predictive analytics and generative AI that staff use for drafting.

The Commission's statement applies the OECD definition: a machine-based system that infers from its inputs how to generate outputs such as predictions, content, recommendations or decisions. It excludes rules-based automation that follows pre-set instructions. That gives a workable test. If a tool writes, predicts, ranks or recommends, treat it as AI. If it only follows fixed rules, it is automation, which still needs oversight but sits outside the Commission's definition.

The Department of Health, Disability and Ageing has raised the same risk. Its February 2026 Tech Talk slides warn that staff could already be using AI even if no tools have been officially rolled out, and ask whether permitted and non-permitted uses are defined, what client data can be used and whether staff have training. The department's June 2026 session then pointed providers to AI.gov.au as a single entry point for government AI guidance. Governa's piece on shadow AI in aged care covers how unapproved tools creep in.

Several sessions at this month's Ageing Australia conference deal with AI governance, and Governa's Ageing Australia National Conference session guide lists them.

A five-step starting point

List every tool that writes, predicts, ranks or recommends, including free tools staff use on their own.
Name an accountable owner for each tool and record who approves new ones.
Write down permitted and non-permitted uses, including which client information may be entered. The AI and automated decision-making policy template gives a starting structure.
Set the points where a person must review or override the output, and check supplier terms on data handling. The article on AI scribe vendor risk covers what to review when a supplier changes shape.
Train staff, keep records of approvals and reviews, and set a review date for each tool.

Related Resources

Common Questions About AI in Aged Care

1. Is there a rule that tells aged care providers how to use AI?

The Commission's statement describes its own use of AI and is not a rule for providers. Your existing duties under the Aged Care Act 2024, the Statement of Rights, the Code of Conduct and the Quality Standards still apply to care that AI supports. Ask your legal adviser how they apply to a specific tool.

2. Do we need an AI register?

Nothing in the Commission's statement requires providers to keep one, but the Commission runs one for itself. A register is the simplest way to answer who uses what, with which data and under whose approval.

3. Is a note-taking or transcription tool AI?

If it listens or reads and then generates a summary, a note or a recommendation, it meets the OECD definition the Commission applies. Treat it as AI, name an owner and set a point where a clinician checks the output before it enters the record.

4. Will the December 2026 privacy change cover our tools?

Only if a tool makes a decision, or does something substantially and directly related to making one, that could significantly affect an individual's rights or interests, and it uses that person's personal information. Map each tool against the three conditions before you update the privacy policy.

5. Who should be accountable for AI in our organisation?

The governing body is accountable for how the organisation governs AI, and each tool needs a named owner who answers for its outcomes. Governa's article on who leads AI governance in aged care covers how to split those roles.

Show Your Working Before Anyone Asks

The Commission has put its own AI controls in public view, so providers can expect the same question to arrive from boards, auditors and families. You do not need a large program to answer it. You need a list of what is in use, a name against each tool, a point where a person decides and a date to review it. Start this week with the list, because nothing else in the process works until you know what is running.

AI in Aged Care: What the Regulator's Own AI Governance Means for Providers