Privacy Act AI Disclosure: What Australian Businesses Must Do by 10 December 2026

Updated October 2026: The OAIC released its final guidance on 30 September 2026, including a fact sheet, a decision-making flowchart, updated APP 1 guidelines and sample privacy policy wording. This guide reflects it.

From 10 December 2026, new Privacy Act AI disclosure rules require many Australian businesses to explain in their privacy policy when software, including AI, is used to make or help make decisions about people. This guide explains what the rules cover, whether they apply to you, where businesses commonly find AI decisions they didn’t know about, and what to update before the deadline.

Key takeaways

  • The rules start on 10 December 2026 and apply to decisions made from that date.
  • They cover any computer program, not just AI, that makes or substantially informs decisions significantly affecting people.
  • The main obligation is to update your privacy policy. You don’t need to notify each person individually.
  • Using a vendor’s software doesn’t move the obligation to the vendor. The disclosure is yours.
  • If you’re unsure whether something is in scope, the OAIC recommends disclosing it.

What are the Privacy Act AI disclosure rules?

Changing in Privacy Act AI Disclosure

The rules come from the Privacy and Other Legislation Amendment Act 2024, which added three new Australian Privacy Principles: APP 1.7, 1.8 and 1.9. In the law they’re called automated decision-making (ADM) transparency obligations. Most businesses call them AI disclosure rules, because AI is where the change will be felt most.

They’re a transparency requirement. The main thing you need to do is update your privacy policy so people can understand when software plays a part in decisions about them.

The rules apply to decisions made from 10 December 2026, even if you collected the personal information before that date. A customer record you collected in 2023 that feeds into an automated decision in 2027 is covered.

Why Privacy Act AI disclosure matters: enforcement is increasing

The OAIC has made privacy policies an enforcement priority. In January 2026, it began its first ever privacy compliance sweep, reviewing the privacy policies of around 60 organisations across six sectors.

A privacy policy that doesn’t meet the Australian Privacy Principles can lead to a compliance notice, an infringement notice or civil penalty proceedings. With the December deadline now set and guidance published, businesses can expect the OAIC to check for these new disclosures too.

There’s also a trust benefit. Customers and job applicants increasingly want to know when AI is involved in decisions about them. A clear disclosure shows you’ve thought about it.

Do the Privacy Act AI disclosure rules apply to your business?

Step 1: Is your business covered by the Privacy Act?

The Privacy Act generally applies to businesses with an annual turnover above $3 million. Some smaller businesses are covered too, for example health service providers and businesses that trade in personal information. Check whether the Act applies to you before assuming it doesn’t.

Step 2: The three-part test

If you’re covered, you need to disclose in your privacy policy if all three of these apply:

  1. You’ve arranged for a computer program to make a decision, or to do something substantially and directly related to making it. This includes software that scores, ranks, filters or recommends, not just systems that decide on their own.
  2. The decision could reasonably be expected to significantly affect someone’s rights or interests. Think approving or declining an application, setting a price, shortlisting job candidates or flagging an account for review.
  3. Personal information about that person is used by the program to make or inform the decision.

A human signing off at the end doesn’t automatically take you out of scope. If the software’s output is a key factor in the decision, it can still count. A decision also includes refusing or failing to make a decision, such as an application that’s automatically left in a queue.

If you’re unsure, the OAIC recommends taking a cautious approach and including the information in your privacy policy.

Examples: where businesses may need to disclose

These are common situations worth checking against the three-part test. Whether each one is in scope depends on how the software is actually used in your business.

Business areaExampleWhy it may be in scope
RecruitmentA hiring platform scores or ranks applicants and only top-ranked candidates are interviewedThe ranking directly shapes who gets a job opportunity
Real estateTenancy software scores rental applications before a property manager reviews themThe score is a key factor in whether someone gets a home
Finance and lendingCredit or affordability tools approve, decline or set limitsThe outcome directly affects access to money
InsuranceSoftware calculates premiums or flags claims for investigationPrice and claim outcomes significantly affect the customer
Retail and onlineDynamic pricing or eligibility rules use customer data to set offersPricing can significantly affect a person’s interests
Customer accountsFraud tools automatically suspend or restrict accountsLosing account access is a significant effect
HealthcareTriage or booking tools prioritise patients based on their informationPriority of care is a significant effect

What usually won’t need disclosing

Many everyday uses of AI are unlikely to meet the test, because they don’t make or substantially inform decisions that significantly affect people. For example:

  • Using Microsoft Copilot or ChatGPT to draft emails, documents or social posts
  • Summarising meeting notes or internal reports
  • Spam filters and basic cybersecurity tools that protect your systems
  • Website analytics that report on traffic trends without making decisions about individuals

These uses still have privacy risks, which we cover below, but they generally aren’t what APP 1.7 is aimed at.

What your privacy policy must disclose

If the test applies, your privacy policy needs to describe:

  • The kinds of personal information used by those computer programs
  • The kinds of decisions made solely by a computer program
  • The kinds of decisions where a computer program does something substantially and directly related to the decision, such as scoring or recommending

You describe kinds of information and decisions, not every individual system or vendor. The disclosure should be clear enough that a person could understand what’s happening and ask you about it.

Example privacy policy wording

Here’s a simplified example of how a business using recruitment software might word its disclosure. It’s for illustration only. Start from the OAIC’s sample wording and have the final version reviewed.

Automated decision-making

We use software to help us assess job applications. The software uses the information in your application, such as your work history, qualifications and responses to screening questions, to score and rank applications. Our recruitment team uses these rankings to decide which applicants to invite to interview. We do not make final hiring decisions solely by software. If you have questions about how your application was assessed, contact us using the details in this policy.

Is Privacy Act AI disclosure only about AI?

No. Despite the name most people use, the rules cover any “computer program”, including traditional rules-based software that has been around for years. In practice, though, AI is where many businesses will find new decisions to disclose, because AI features are now built into so many everyday platforms.

Using a third-party tool doesn’t move the obligation to the vendor. If your business arranged for the software to be used in your decisions, the disclosure is yours to make.

How to find AI decisions hiding in your existing software

The hardest part of Privacy Act AI disclosure isn’t writing the policy. It’s knowing where automated decisions are happening. Many platforms have added AI features through updates, sometimes switched on by default.

Check these places first:

  • CRM and sales platforms: lead scoring, customer segmentation and automated follow-up rules
  • Recruitment and HR platforms: applicant screening, ranking and automated rejections
  • Finance and accounting tools: credit checks, payment risk scoring and automatic account holds
  • Industry platforms: property management, practice management and booking systems with built-in scoring or prioritisation
  • Microsoft 365 and Copilot: usually low risk for drafting, but check whether any Copilot agents or automations are being used in decisions about customers or staff
  • Custom automations: workflows built in tools like Power Automate or Zapier that approve, decline or route requests automatically

Questions to ask your software vendors

  1. Does your product use personal information to make, score, rank or recommend decisions about individuals?
  2. Which features do this, and are they switched on for our account?
  3. What kinds of personal information does each feature use?
  4. Can we switch these features off, or add human review?
  5. Can you give us a plain-English summary we can use in our privacy policy?

The OAIC expects vendors to give their customers clear, high-level information so businesses can make the right disclosures, so these are reasonable questions to ask.

Do I need to label AI-generated content?

Not under these rules. Privacy Act AI disclosure is about your privacy policy. It doesn’t require you to label AI-written content or notify each person individually every time software is involved in a decision. Other laws, such as Australian Consumer Law, can still apply if AI use could mislead customers.

Putting personal information into AI tools: the secondary use question

Separate from the December rules, the Privacy Act already limits how you use and disclose personal information. Under APP 6, you can generally only use or disclose personal information for the purpose you collected it for, unless an exception applies, such as consent or a related purpose the person would reasonably expect.

That matters when staff paste customer or employee information into AI tools like chatbots or copilots. The OAIC’s guidance on commercially available AI products recommends caution, especially with sensitive information and public AI tools. A clear AI usage policy for staff is the simplest way to manage this risk, alongside business-grade AI tools that don’t use your data to train public models.

Privacy Act AI disclosure checklist: 5 steps before 10 December

  1. List your systems. Identify software, including AI features inside platforms you already use, that uses personal information in decisions about customers, applicants or staff.
  2. Apply the three-part test to each one. Use the OAIC’s flowchart to help, and record your reasoning.
  3. Ask your vendors what their tools do with personal information and how outputs are used.
  4. Update your privacy policy with the kinds of personal information and decisions involved. Use the OAIC’s sample wording as a starting point.
  5. Set up a review process, so new software and AI features are checked before they’re switched on, and your privacy policy stays accurate.

What’s coming next

The December rules are part of the first tranche of Privacy Act reforms. The Government has indicated that a second tranche is likely to give individuals a new right to request meaningful information about how substantially automated decisions with significant effects are made.

That means the inventory you build now won’t just help with your privacy policy. It will put you in a much stronger position for the next round of reforms.

This article is general information, not legal advice. For advice on your specific situation, speak to a privacy lawyer.

Need help getting ready?

XCD IT can help you list the software and AI tools your business uses, work out where personal information flows, and put an AI usage policy in place before 10 December. Call us on 07 3085 7447 or read our AI & Privacy Act compliance checklist.

Frequently Asked Questions About Privacy Act AI Disclosure

10 December 2026. They apply to decisions made from that date, including decisions that use personal information collected earlier.

Yes. On 30 September 2026, the OAIC released a fact sheet, a decision-making flowchart, updated APP 1 guidelines and sample privacy policy wording.

No. The requirement is to include the information in your privacy policy.

Not automatically. It depends on whether the tool is used to make, or substantially inform, decisions that significantly affect people using their personal information. Drafting emails usually wouldn’t. Screening job applications could.

Possibly. Human review doesn’t automatically take a decision out of scope. If the software’s output is a key factor in the final decision, it can still be covered.

You may be exempt, but not always. Some small businesses, such as health service providers, are covered regardless of turnover. Check before relying on the exemption.

The OAIC can issue compliance notices and infringement notices, or take civil penalty action, where a privacy policy doesn’t meet the Australian Privacy Principles.

Jack Richards

Latest Posts