AI in workplaces
AI in Staff Decisions and Your Privacy Policy from December
From 10 December 2026, APP entities must say in their privacy policy when a computer program, including AI, makes or does something substantially and directly related to a significant decision about a person. The policy must also say what kinds of personal information the program uses. If your team uses AI in rostering, recruitment, eligibility or bonuses, you may need to tell whoever looks after your privacy policy.
What changes on 10 December 2026?
The OAIC announced new resources on 30 September 2026. They explain a new transparency obligation under APP 1.7 to 1.9 of the Privacy Act.
An APP entity must include the extra information in its privacy policy if three things are true.
- It has arranged for a computer program to make a decision, or do something substantially and directly related to making one.
- The decision could reasonably be expected to significantly affect a person's rights or interests.
- Personal information about the person is used to run the program.
The policy must then cover the kinds of personal information the programs use. It must also cover the kinds of decisions made only by the programs, and the kinds where a program does a related step.
Which decisions count as significant?
The OAIC fact sheet says the effect must be more than trivial. It must have the potential to considerably influence a person's circumstances or outcomes.
A computer program is not only AI. The fact sheet includes rule-based processes, machine learning, standard software and generative AI such as chatbots. A decision can also be a choice not to act, such as refusing something. It can help a person or disadvantage them.
The fact sheet gives examples that touch staff and job seekers. One is recruitment software that screens video interviews. Another is a generative AI report used to inform staff bonuses.
Say your organisation uses a spreadsheet tool to rank rostering requests, or software to sort job applications. Those are the kinds of uses worth raising, because they can affect someone's work or income.
Does a human check take you out of scope?
Not automatically. The OAIC's APP 1 Guidelines say a decision may be within scope even where the program's output is subject to human review. A program can also recommend a decision to a person. It can still count when the decision maker relies on it or is significantly influenced by it.
The guidelines say a program used only for other purposes, such as word processing, is not captured.
The guidelines also say the obligation stays with the entity using the personal information, even when a third party supplies the software. Buying a tool does not hand the duty to the vendor.
The fact sheet's bonus example points to something useful. It notes that steps such as further review and documenting why a manager diverged from the AI output could indicate the output was not substantially related to the decision. That is a good reason to keep records.
What should staff and managers do now?
This is a transparency rule, so most of the work is knowing and telling. These steps are practical suggestions, not extra legal requirements.
- List the tools. Write down which programs your team uses when deciding things about people. Include free tools, add-ons and features switched on inside software you already have.
- Note the personal information. For each tool, record what kinds of personal information go in. Examples are work history, availability, leave records or health details. The fact sheet says it should be clear when more sensitive information, such as health information, is used.
- Tell the policy owner. Pass the list to whoever owns your privacy policy. That is often a privacy officer, a manager or a legal contact. If you are unsure who it is, ask.
- Keep a real human review. Someone should read the output, check it against the facts, and be willing to change it.
- Write down departures. When you decide differently from the AI output, note what you changed and why. A short file note is enough.
- Ask your vendor. The fact sheet says vendors should provide high-level information about their software so the entity can complete its disclosures.
If you are not sure whether a use is in scope, the fact sheet says to take a cautious approach and include the information. Raise it with the policy owner rather than deciding alone.
Who has to follow this?
The obligation applies to APP entities. The OAIC's announcement and fact sheet use that term, and they do not give your organisation a yes or no answer here. Your privacy officer or legal adviser is the right person to confirm whether your organisation is covered.
Where can you learn more?
If you want a practical starting point for using AI at work, the course Responsible AI for Australian Workplaces covers the basics. Teams can also look at AI Work Ready for organisations.
This is general information, not legal advice. Check your organisation's policy.
Sources
Learn the habit, free
Responsible AI for Australian Workplaces teaches what never goes into an AI tool, how to ask well and how to check the answer. Short online lessons, free for individuals.