AI in healthcare teams
Automated decision-making and your privacy policy: does yours need updating?
Only if a tool you use meets the conditions the Office of the Australian Information Commissioner (OAIC) describes, and the obligation starts on 10 December 2026. From that date, an organisation covered by the Australian Privacy Principles must say in its privacy policy when it has arranged for a computer program to make a decision, or to do something substantially and directly related to making one, using a person's personal information, where the decision could reasonably be expected to significantly affect that person's rights or interests. If all of those conditions are met, the policy must name the kinds of personal information the program uses and the kinds of decisions it makes or helps make. Whether a particular tool meets them is for you and your privacy adviser to decide, and the OAIC says that if you are in doubt, take a cautious approach.
What does the new transparency rule say?
It adds three conditions to Australian Privacy Principle 1, the principle that requires a privacy policy. The OAIC calls it the APP 1.7–1.9 transparency obligation. The OAIC's updated APP 1 Guidelines say that from 10 December 2026, APP 1.7 requires the extra information when:
- the organisation has arranged for a computer program to make a decision, or to do a thing that is substantially and directly related to making a decision;
- the decision could reasonably be expected to significantly affect the rights or interests of an individual; and
- personal information about that individual is used in the operation of the program to make the decision or do the thing.
All three have to apply. The OAIC's flowchart walks through them one question at a time, and reaches "in scope" only when every answer is yes.
The OAIC reads "decision" broadly. The Guidelines say it includes refusing or failing to make a decision or do a thing, and that a decision can affect someone whether the effect is good or bad for them.
Does a person reviewing or making the final decision take it out of scope?
Not on its own. The Guidelines say a decision may be in scope even where the program's output does not replace the whole decision-making process, or is subject to human review. They say a program does not need to make the final decision, or be the only decision-maker, and that its output can still be substantially and directly related to a decision where the person deciding relies on it or is significantly influenced by it. In one of their examples, a panel makes the final decision on scholarship applications and can disregard a program's ranking. The OAIC still treats the ranking as substantially and directly related to the decision, because the panel considers it.
The fact sheet says the OAIC considers machine learning or generative AI outputs used to make decisions that significantly affect people's rights or interests "would generally fall within scope" of the obligation, "unless subject to extensive human oversight and control". None of the OAIC's resources on the obligation defines "extensive". The self-check below helps you ask where your team's use of each tool sits.
Who does it apply to?
APP entities: the organisations the Australian Privacy Principles cover. The flowchart's first question is whether the organisation is an APP entity at all. The OAIC's resources on this obligation do not separately mention small businesses or not-for-profits; they apply it to APP entities. Whether a small practice, not-for-profit or NDIS provider is one is a separate question, and this guide does not answer it.
If your organisation is an APP entity, the obligation sits with you, not with the software vendor. The OAIC's fact sheet says it "generally expects" the transparency obligation to remain with the organisation that uses personal information to make the decision. It also says providers of third-party software should give their customers clear, high-level information about how the software can be used to make decisions, so the customer can describe it in its privacy policy.
Buying a product can count. The fact sheet lists, among the ways an organisation may have "arranged for" a program to make or inform a decision: procuring a third-party program, configuring or customising off-the-shelf software, and relying on a program's advisory outputs.
Self-check: questions to ask about each tool your team uses
- Does the tool use personal information about patients, participants, staff or anyone else? The flowchart says a computer program is only in scope to the extent personal information is used to make or inform a decision.
- Did we arrange for it? Did we build it, buy it, configure it, or come to rely on what it suggests? The fact sheet lists all of these as ways an organisation may have "arranged for" a program, and the Guidelines say an organisation may still be covered when another company provides the program under contract.
- Is it a "computer program"? Probably. The Guidelines say the term is read broadly and covers rule-based processes, AI and machine learning, everyday software, apps and word-processing tools, and generative AI including chatbots. This is only the first step: being a computer program does not put a tool in scope on its own.
- Does it make a decision, or feed one? Does it decide, recommend, rank, score, prioritise or flag something a person then acts on, including deciding not to act? The fact sheet says a program may be substantially and directly related to a decision if it advises of an appropriate outcome, determines the outcome or "in any other way influences the outcome".
- How much does the decision rest on it? The Guidelines list factors to weigh: how heavily staff rely on the output, whether staff can and do override it, whether the output is advisory or determinative, whether it explains what it did and why, and how deeply it is built into the workflow.
- Or does it only record a decision a person has already made? The Guidelines say that where a person makes the decision and uses a program for purposes other than facilitating it, "such as using a word processing program to document a decision", this will not be captured. Ask honestly which side of that line your use falls on.
- Could the decision significantly affect someone? The OAIC says "significantly" means more than trivial, with the potential to considerably influence the person's circumstances or outcomes. Its list of significant rights or interests includes access to healthcare, emergency or aged care services (including "in home" services) and access to physical assistance or support.
- Are the people affected more vulnerable? The Guidelines say a decision's effect on, for example, a child or a person with a disability may be considered significant compared with its effect on others.
- Can the vendor tell us, in writing, how the tool makes or informs decisions? You will need this to describe the tool in your privacy policy, and the fact sheet expects software providers to supply it.
- If we are still unsure, who decides? Write down your answers and take them to your privacy adviser. The OAIC says organisations "in doubt" should take a cautious approach and include the APP 1.8 information in their privacy policy.
The self-check helps you ask the question about each tool. It does not answer it: the result depends on how your team uses the tool, and this guide does not say whether any product is in or out of scope.
Which tools in a practice or NDIS service might this touch?
The OAIC does not name AI scribes, rostering systems or practice software in these resources, so there is no official list to check a tool against. What it does give are examples close to health and disability services. The fact sheet's list of decisions that would generally be in scope includes "computer programs used to prioritise the provision of health or disability services to individuals". One of its worked examples is a health and aged care provider whose spreadsheet formula scores and ranks clients, and whose staff use the ranking to decide whom to contact. The OAIC treats that as in scope, including where the ranking is used to decide whom not to contact.
The examples also reach staff decisions. The fact sheet's list includes recruitment software used to make hiring decisions, and AI-generated reports used to rank employee performance or decide promotions or pay. The Guidelines give the example of an employer that permits or directs employees to use an AI chat tool to draft performance assessments that determine promotion decisions.
The RACGP's AI scribes page says the new provisions "may affect the use of AI scribes". Neither the RACGP nor the OAIC says a particular scribe or tool is in scope. Work through the self-check for each tool and take the answers to your adviser.
What must the privacy policy say?
Three things, under APP 1.8, if all the conditions are met:
- the kinds of personal information used in the operation of those computer programs;
- the kinds of decisions made solely by the operation of those programs; and
- the kinds of decisions for which a thing substantially and directly related to making the decision is done by those programs.
The OAIC asks for meaningful information, not excessive detail. The Guidelines say organisations must strike a balance between giving enough for people to understand their use of a computer program and avoiding detail that obscures the point. They say the information should be in plain language, tailored, grouped logically, and framed so that the information and the decision can be challenged or contested. Organisations may group kinds of decisions and kinds of personal information, but the Guidelines say a privacy policy should clearly say where more sensitive information, such as health information, is used.
Commercial-in-confidence information about the program, such as commercially sensitive information or trade secrets, does not have to be included. The fact sheet ends with worked examples, including two sample privacy policy sections written by the OAIC. Read those for the level of detail it has in mind, and have your adviser check your own wording.
Does the vendor's answer settle it?
No. A vendor can tell you how its product works, but the Guidelines say organisations should "actively identify, assess and keep oversight over" the third-party programs they use to make decisions, both during and after procurement. They say organisations should understand how a program is being used to make or assist decisions and what kinds of decisions are being made, and that contracts should be clear about which party keeps the final decision-making capability or judgement.
A practice or service AI policy is a natural place to record which tools you use, who approved them and what you concluded about each one. For the wider picture, see the AI in healthcare teams hub.
What's changing?
- 10 December 2026: the obligation starts. The OAIC's news item of 30 September 2026 says that from this date, APP entities must include information in their privacy policy where they have arranged for a computer program to make or inform decisions that significantly affect individuals' rights or interests. It says the obligation was introduced into APP 1 by the Privacy and Other Legislation Amendment Act 2024.
- 30 September 2026: OAIC resources published. The same news item announced the fact sheet, a supplementary fact sheet for government agencies and the flowchart, and an update to the APP 1 Guidelines. The Guidelines chapter shows "Updated: 30 September 2026" and version 2.0.
- RACGP. The RACGP's AI scribes page says the new provisions will extend the Australian Privacy Principles to certain automated decision-making processes, and that this "may affect the use of AI scribes".
Check the OAIC pages above for changes before you rely on this summary.
This page is general information, not legal advice. It summarises the OAIC's resources on the APP 1.7 to 1.9 transparency obligation and does not say whether any tool, privacy policy or organisation meets it. Confirm your obligations with the OAIC or your own privacy adviser or lawyer.
How we help: Make informed decisions about AI
Sources
- OAIC, Chapter 1: APP 1 Open and transparent management of personal information (APP Guidelines) (Updated 30 Sep 2026 (version 2.0))
- OAIC, APP 1.7-1.9 Transparency Obligation: Fact Sheet (September 2026)
- OAIC, APP 1.7-1.9 Transparency Obligation Flowchart (Undated)
- OAIC, New resources on transparency for use of AI and automated decision-making (30 Sep 2026)
- RACGP, Artificial intelligence (AI) scribes (Header 29 Jun 2026; publication info: first published Aug 2024, updated Oct 2025)
A useful place to start
Better technology.
Better care, together.
Tell us about your team and what you want to make better. We’ll find a useful place to begin.