Using AI in hiring lawfully in Australia
Reviewed by Mellow Editorial Team, HR & payroll content team
Using AI tools to screen CVs, rank candidates or conduct video interviews can save significant time — but in Australia those tools are subject to existing discrimination, privacy and employment laws that were written well before AI existed.
What the law already covers
Australia does not yet have AI-specific hiring legislation, but several existing frameworks apply directly.
The Fair Work Act 2009, state and federal anti-discrimination laws (including the Age Discrimination Act, Disability Discrimination Act, Racial Discrimination Act and Sex Discrimination Act), and the Privacy Act 1988 all govern how you collect information about candidates and how you make decisions about them. The fact that a decision was made or assisted by an algorithm does not create a legal exemption. If an AI tool produces an outcome that would be unlawful if a human produced it, it is still unlawful.
The Australian Human Rights Commission has flagged automated decision-making in hiring as a priority area for scrutiny. The government's 2024 voluntary AI Safety Standard also signals a clear direction of travel toward mandatory guardrails, particularly for high-stakes decisions — and hiring is explicitly treated as high-stakes.
Discrimination risk is real and specific
AI screening tools are trained on historical data. If that data reflects past hiring patterns — for example, favouring candidates from certain universities, certain suburbs, or certain demographic profiles — the model will replicate and potentially amplify those patterns.
Under Australian anti-discrimination law, both direct and indirect discrimination are prohibited. Indirect discrimination occurs when a requirement or condition that appears neutral disproportionately disadvantages a group with a protected attribute (age, race, sex, disability, pregnancy, and others). An AI tool that systematically down-ranks candidates from a particular background could constitute indirect discrimination even if the tool never asks about that characteristic explicitly.
Practical steps to reduce this risk:
- Ask your vendor to explain what training data was used and whether it was audited for bias before deployment.
- Run a sample of AI-screened outcomes against your own demographic data if you collect it, and look for unexplained patterns.
- Never make a final hiring decision based solely on an AI score. Keep a human in the loop for every shortlisting or rejection decision.
- Document your rationale for the criteria you set in the tool. If challenged, you need to be able to explain why those criteria are relevant to the role.
Privacy obligations when using AI hiring tools
The Privacy Act 1988 and the Australian Privacy Principles (APPs) apply whenever you collect personal information about candidates. Several obligations are directly relevant to AI hiring tools.
Collection notice. You must tell candidates what information you are collecting, why, and how it will be used or disclosed. If you are using a third-party AI platform that processes or stores candidate data, that needs to be disclosed.
Sensitive information. Some AI tools analyse voice, video or facial data. Health, genetic and biometric information is classified as sensitive information under the APPs and attracts stricter handling rules, including a requirement for consent in most circumstances.
Data minimisation. Collect only what you genuinely need for the hiring decision. An AI tool that ingests a candidate's entire social media footprint is likely collecting more than is reasonably necessary.
Cross-border disclosure. If the AI vendor stores or processes data outside Australia, the APP 8 cross-border disclosure obligations apply. Your privacy policy should reflect this, and you should confirm the vendor's data storage arrangements before signing a contract.
What to ask your AI vendor before you sign
Vendor due diligence is not optional. Before deploying any AI hiring tool, get clear answers on:
- Where candidate data is stored and processed, and which countries are involved.
- How long candidate data is retained and whether candidates can request deletion.
- What the model was trained on and whether bias testing has been conducted.
- Whether the tool produces an explainable output (a reason, not just a score) that you could reproduce in a dispute.
- What the vendor's process is if the tool produces a discriminatory outcome.
If a vendor cannot answer these questions clearly, that is itself useful information.
Building a defensible process
The goal is not to avoid AI in hiring — the efficiency gains are genuine. The goal is to use it in a way you could defend to a candidate, a tribunal or a regulator.
A defensible process typically looks like this: set criteria tied explicitly to the role requirements, apply them consistently, use the AI output as one input rather than the final decision, keep records of how shortlisting decisions were made, and give candidates a clear avenue to raise concerns.
Candidates do not currently have a statutory right to an explanation of an automated hiring decision in Australia, but that may change. Building explainability into your process now is straightforward and protects you if the law moves in that direction — which, based on current policy signals, it is likely to do.
---
Run HR and payroll in Australia with Mellow
Mellow brings HR, payroll and 12 AI agents into one platform — built to handle Australia properly, with payroll included, from £4 per employee per month. The AI agents don't just answer questions; they generate contracts, run cost estimates and draft letters for you.
[Start a free trial →](/register)