New automated decision-making privacy rules start on 10 December 2026. From that date, many Australian businesses must explain in their privacy policy when software or AI uses personal information to make, or help make, decisions that significantly affect people.
The change sounds narrow. However, the regulator’s guidance shows it can capture everyday tools, such as recruitment software, credit checks and even spreadsheets with scoring formulas. This guide explains what the rules require, which tools may be caught and what to do before the deadline.
Updated 19 November 2026
What’s changed
- New disclosure rules start 10 December 2026. They add new requirements, APPs 1.7 to 1.9, to the Australian Privacy Principles.
- OAIC guidance released (30 Sep 2026). The Office of the Australian Information Commissioner published guidance on the new obligation.
- It’s broader than “fully automated”. Decision-support tools, embedded features and simple software can be caught.
- Existing data counts. The rules apply to decisions made after 10 December, even if the information was collected earlier.
What the automated decision-making privacy rules require
Under the new rules, your privacy policy needs extra information when all three of the following tests are met.
A program is involved
You have arranged for a computer program to make a decision, or to do something substantially and directly related to making one.
The decision is significant
The decision could reasonably be expected to significantly affect a person’s rights or interests.
Personal information is used
The program uses that person’s personal information to do its job.
What your privacy policy must disclose
If all three tests are met, your privacy policy must describe:
- the kinds of personal information those programs use;
- the kinds of decisions made solely by those programs; and
- the kinds of decisions where a program solely does a step that is substantially and directly related to the decision.
For example, decisions that affect someone’s rights under a contract, or their access to a significant service, are likely to be significant. Similarly, the guidance points to areas such as finance, housing, health and employment.
Which tools could be caught?
The guidance makes clear that the rules are not limited to advanced AI. Here are common examples worth checking.
Recruitment tools
Software that screens, scores or ranks job applicants. Targeted job advertising can also be in scope.
Credit and payment checks
Tools that decide credit limits, payment terms or whether to offer an account.
Eligibility and pricing
Systems that decide who qualifies for a service, a discount or a price.
Chatbots and triage
Generative AI or chatbots that decide how requests are handled or prioritised.
Fraud and identity checks
Automated checks that can block an account, payment or application.
Spreadsheets and rules
Even simple formulas that score or rank people can count if they drive the decision.
A person signing off may not be enough. According to the guidance, human review doesn’t automatically take a tool out of scope. If the program’s output is a key factor in the decision, it is likely to be caught. In addition, using a vendor’s product doesn’t shift the obligation to the vendor.
Does this apply to your business?
The rules apply to organisations covered by the Privacy Act. Generally, that means businesses with annual turnover over $3 million, plus some smaller businesses, such as health service providers.
If you’re covered, the next question is whether you use software or AI in significant decisions about people. Many businesses do, often without realising it, because these features are built into everyday platforms for HR, finance and customer service.
Six steps to take before 10 December
With only weeks to go, focus on these steps first.
List your decisions
Write down the decisions you make about customers, staff and job applicants that matter to them.
Find the software
Next, identify which systems, apps and AI tools play a part in each decision.
Apply the three tests
Then check each one against the three tests above, and note why it is in or out of scope.
Ask your vendors
Request plain-English details of how their tools use personal information.
Update your policy
Add the required information to your privacy policy, ideally with advice from a privacy professional.
Keep it current
Finally, review the list whenever you add new software or AI features.
How CAG helps with automated decision-making privacy
Our team can help you find the software and AI tools that touch personal information, so your privacy adviser can update your policy with confidence. First, we document the systems and processes involved. Then we help you understand how each tool reaches its outputs and where people stay in the loop.
In addition, our AI consulting and automation services include AI use policies and workflow automation designed with clear human checkpoints. For practical examples, read our guides to business process automation and AI in the workplace.
We’re not lawyers, so we work alongside your legal or privacy adviser on the policy wording itself.
Automated decisions
Not sure which of your tools are caught?
Our team will help you map the software and AI tools that use personal information, so you can update your privacy policy before 10 December.
Automated decisions and privacy: frequently asked questions
The new rules
When do the automated decision rules start?
They start on 10 December 2026. From then, your privacy policy needs to describe relevant automated decisions, including those based on information you collected before that date.
Do the rules ban automated decisions?
No. The rules are about transparency. You can keep using software and AI to make or support decisions, but your privacy policy must explain it.
Does human review take a tool out of scope?
Not necessarily. If a program’s output is a key factor in the decision, it can still be caught, even when a person signs off.
Getting ready
Does this apply to small businesses?
It applies to organisations covered by the Privacy Act. Generally, that means businesses with annual turnover over $3 million, plus some smaller ones, such as health service providers.
We use a vendor’s software. Is that their problem?
No. The obligation sits with the business using the tool, not the vendor. So you may need to ask suppliers how their tools work.
Do we need a lawyer to update our privacy policy?
For many businesses, it’s wise to have a privacy lawyer or adviser review the changes. We can help you map the systems and tools involved first.
Sources and further reading
- OAIC — APP guidelines, Chapter 1: Open and transparent management of personal information
- Allens — Mandatory automated decision-making disclosure requirements are coming
- Gilbert + Tobin — Automated decision-making transparency under the Privacy Act
This article is general information, not legal advice. Get advice from a privacy professional about your specific obligations.

About the author
Gordon Murdoch
Director (Policy, Compliance and Training), The Computing Australia Group
Gordon leads CAG’s policy, compliance and training work. He helps organisations turn security requirements into practical policies their teams can actually follow, and drives projects that use AI and other tools to lift efficiency and service quality. His view is simple: compliance should make a business stronger, not just tick a box.