AI in recruitment is lawful in Australia, and it works best when it writes, summarises, schedules, and ranks while a person makes every decision about a candidate. Before switching it on, check for bias, tell candidates, protect their personal information, and update your privacy policy for the automated decision rules starting 10 December 2026.
Key takeaways
- The Australian Public Service Commission’s AI recruitment principles state that AI must not make the final decision, and agencies were expected to apply them by 1 June 2026.
- Those principles bind government agencies only, but a private employer can borrow them as a ready-made benchmark.
- From 10 December 2026, privacy policies must explain computer programs that make, or substantially inform, decisions that significantly affect people.
- According to the Merit Protection Commissioner, 15 of 66 APS agencies surveyed in 2022 (23 per cent) had used AI-assisted or automated recruitment tools.
- The employee records exemption does not cover unsuccessful job applicants, so their information falls under the Australian Privacy Principles.
Where is AI in recruitment used today?
AI in recruitment now appears at five points in the hiring process: writing job ads, screening and ranking applicants, assessing candidates by chat or video, scheduling interviews, and sourcing people who have not applied. Each point carries a different level of risk, because some uses touch documents while others touch decisions about people.
Most Australian guidance on the subject was written for government agencies. However, the same tools are now sold to private employers of 50 to 1,000 people, usually inside an applicant tracking system or HRIS. The five uses look like this in practice:
- Writing job ads: AI turns a position description into an advertisement, a set of screening questions, and a batch of candidate emails.
- Screening and ranking: AI reads each application, scores it against the criteria for the role, and orders the list.
- Assessments by chat or video: candidates answer set questions in their own time, and AI scores or summarises the answers.
- Scheduling: AI books interviews, sends reminders, and answers routine questions from candidates.
- Sourcing: AI searches past applicants and outside profiles for people who fit the role.
Adoption started earlier than many employers assume. According to the Merit Protection Commissioner, 15 of 66 APS agencies surveyed in 2022 had used AI-assisted or automated recruitment tools. That is 23 per cent, measured over the previous 12 months. Saving time was the benefit those agencies cited most often.
Where does AI in recruitment help, and what should you avoid?
AI helps most with the writing, summarising, and scheduling around a hiring decision, and it can rank applicants when a person checks the ranking. Avoid any use where software alone rejects a candidate, infers personality or emotion from video, or handles candidate information in a public chatbot.
The NSW Office of the Public Service Commissioner gives a useful list of sensible uses. It names drafting recruitment documents and communications, finding key themes in candidate responses or interview notes, answering applicant enquiries by chatbot, and scheduling interviews. Similarly, the Australian Public Service Commission suggests starting small, for instance with interview notes or a first draft of a selection report.
Both regulators draw the line at the decision. The NSW guidance says AI should not be used to filter out applications or make other decisions about candidates. In addition, the Merit Protection Commissioner recommends minimising hard screening decisions, where one AI-assisted assessment screens a candidate out.
Suppose a retailer advertises a store manager role and receives 300 applications in a week. A ranked list tells the hiring manager who to read first, which saves days. It should not decide who never gets read, so someone still reviews the lower half before anyone is declined.
What are the real risks of AI in the recruitment process?
Five risks matter most: bias that leads to discrimination, rejection by software with no person involved, mishandled candidate information, candidates who were never told AI was used, and candidates using AI themselves. Each one has a practical control, and none of them requires a ban on the technology.
AI bias in recruitment software
AI can reproduce the bias of its developers and its training data. The Merit Protection Commissioner warns that a tool tested on a narrow group may disadvantage everyone outside it. Its example is memorable: if every successful candidate in the training interviews coughed, the tool may treat coughing as a sign of suitability.
Video tools carry a particular risk. The NSW guidance lists video interview tools that make inferences from emotion, personality, or image, including skin colour, facial hair, or head coverings. It also warns against relying on AI to create equity or prevent discrimination.
Diversity Council Australia reached a similar view after a three-year project with Hudson RPO and Monash University. Its employer guidelines set out a five-step process to help employers tread carefully around diversity and inclusion risks. The steps spell T.R.E.A.D.: Team Up, Reflect, Educate, Acquire, Decide.
Automated rejection without a person
The NSW guidance names automation bias, where hiring teams rely on AI outputs without questioning them. That risk grows when the tool is wrong, as the Merit Protection Commissioner’s survey shows. Among APS agencies that stopped using AI tools, most said the tools were not working as intended.
The control is review by a person. For shortlisting, the Australian Public Service Commission describes a moderation step or a panel review of sampled applications. The sample should be large enough to show that the tool and the panel reach the same view.
Candidates’ personal information
Applications hold personal information, and sometimes health information. According to the OAIC, privacy obligations cover personal information entered into an AI system and the output it generates. Inferred or generated information about an identifiable person also counts, so an AI score for a named candidate is covered.
As best practice, the OAIC recommends not entering personal information into publicly available generative AI tools. A recruiter who pastes CVs into a free chatbot is doing exactly that. So name the approved tools, and keep candidate information inside them.
Transparency to candidates
Candidates who discover AI was used after the event lose trust in the process. The OAIC says businesses should update privacy policies and notices with clear information about their use of AI. It also says public-facing AI tools, such as chatbots, should be clearly identified as AI.
Candidates using AI themselves
More candidates now use tools such as ChatGPT to write cover letters and answers to selection criteria. The NSW guidance treats the response as a choice for each agency, stated clearly when the role is advertised. Whatever the choice, pair written answers with interviews, work samples, and reference checks, so you assess the person rather than the output.
What does Australian law say about AI in recruitment?
Australia has no law written specifically for AI in recruitment. Instead, existing discrimination and privacy laws apply to hiring however the decision is made. In addition, a new Privacy Act transparency rule for automated decisions starts on 10 December 2026, and it reaches computer programs that substantially inform a decision as well as those that make one.
Discrimination law
The Australian Human Rights Commission lists four federal discrimination Acts. They are the Age Discrimination Act 2004, the Disability Discrimination Act 1992, the Racial Discrimination Act 1975, and the Sex Discrimination Act 1984. State and territory laws sit beside them, and the Commission says employers must comply with all of them.
Diversity Council Australia makes the link to AI directly. In its words, existing anti-discrimination legislation covers recruitment, and through that also AI recruitment. So a tool that screens out older applicants, or candidates with a disability, creates the same exposure as a person doing it.
Privacy law
The OAIC states that the Privacy Act applies to all uses of AI involving personal information. Many employers assume the employee records exemption covers recruitment, but it does not. According to the OAIC, the exemption does not cover prospective employees who are not hired, such as unsuccessful applicants.
The new transparency rule was introduced by the Privacy and Other Legislation Amendment Act 2024. From 10 December 2026, it applies where an organisation has arranged for a computer program to make a decision. It also applies where the program does something substantially and directly related to making that decision. The decision must be one that could reasonably be expected to significantly affect a person’s rights or interests, and the program must use their personal information.
Where the rule applies, the privacy policy must describe three things. It must cover the kinds of personal information used, the kinds of decisions made solely by computer programs, and the kinds of decisions those programs substantially assist. Whether a particular screening tool meets the test depends on how you use it, so take advice before the date rather than after.
What can private employers borrow from the public service principles?
The Australian Public Service Commission’s Principles for agency use of AI in Recruitment bind APS agencies only, with implementation expected by 1 June 2026. However, a private employer can adopt them voluntarily as a benchmark. The core rule is simple: AI tools must not make the final decision in recruitment outcomes.
The Commission published the principles in April 2026, under five headings:
- Human-centred oversight and decision-making: AI assists human judgement and does not replace it, particularly at critical stages.
- Governance, accountability, and risk: a written policy says where AI can and cannot be used, and who is accountable.
- Transparency and explainability: job advertisements tell candidates how, when, and which AI tools will be used.
- Fairness, bias mitigation, and merit: tools must be accessible to every candidate, and they are regularly assessed for accuracy and bias.
- Continuous capability building and adaptability: staff are trained before the tools are used, and practice changes as the tools do.
Two further points travel well. First, an agency stays responsible for AI outputs even when a contracted recruitment provider runs the tool. Second, every recruitment decision must be explainable and defensible. The NSW Commissioner’s Spotlight of July 2025 goes further for NSW agencies, because it says AI should not filter out applications at all.
Neither document applies to a private business. Both show what a regulator regards as careful practice, and that is a useful standard to hold. It gives you an answer when a candidate, a director, or a lawyer asks how a decision was made.
What rules should you set before switching AI on?
Set six rules before AI touches a live role. A person makes every decision, candidates are told, only approved tools see candidate information, each tool is checked for bias, the privacy policy is current, and the advertisement covers candidate use of AI. Write them into your recruitment policy.
Each rule answers one of those risks:
- A person decides: nobody is rejected, shortlisted, or hired until a named person has reviewed the evidence and made the call.
- Candidates are told: the advertisement says how and when AI is used, before any assessment begins.
- Approved tools only: name the tools recruiters may use, and ban candidate information in public chatbots.
- Bias is checked: ask how the tool was tested, then review outcomes regularly for patterns of exclusion.
- The privacy policy is current: update it for AI use, and for the automated decision rule from 10 December 2026.
- Candidate AI use is stated: say whether applicants may use AI, and assess with more than written answers.
Then test the rules on one role before a wider rollout. The Merit Protection Commissioner suggests running a new tool alongside your existing process, to see whether its results are reliable. Compare the AI ranking with the panel’s view, note where they differ, and adjust before the next role opens.
Which AI tool is best for recruitment, and what should you ask a vendor?
No single AI tool is best for recruitment, because the right choice depends on hiring volume, the roles involved, and the systems already in place. The better test is how a vendor answers a short set of questions about decisions, bias, candidate notice, and data. Clear answers matter more than a long feature list.
Seven questions separate careful vendors from the rest:
- Who decides to reject a candidate, and can the tool reject anyone without a person?
- What does the tool score, and can we see the reasoning behind each score?
- Has it been tested for bias with Australian candidates, including people with disability and First Nations people?
- Does it analyse faces, voices, or emotion in video answers?
- What are candidates told about AI, and when do they agree to it?
- Where is candidate information stored, who can access it, and is it used to train models?
- What independent evidence shows the tool measures what it claims to measure?
The last question matters because, as the Merit Protection Commissioner notes, much of the readily available research is sponsored by the providers themselves. For the broader due diligence, the same approach applies to questions for HR software vendors about AI.
Worknice’s recruitment software is one example of the pattern to look for. AI assesses every applicant, and a person on your team makes each hiring decision.
So Worknice suits Australian organisations of 50 to 1,000 people wanting recruitment, onboarding, and HR records on one platform. A recruitment agency, or an employer that needs a vendor holding formal security certification, should look elsewhere. Worknice follows ISO security standards and is not certified. The comparison of AI recruitment software in Australia covers the wider field.
Frequently asked questions
Is it legal to use AI in recruitment in Australia?
Yes. No Australian law bans AI in recruitment, but discrimination and privacy laws apply to every hiring decision however it is made. From 10 December 2026, privacy policies must also explain computer programs that make, or substantially inform, decisions that significantly affect people. Take advice on how these rules apply to your own process.
Can AI reject job applicants automatically?
Software can be set up to do it, but it is the practice regulators warn against most. The Australian Public Service Commission says AI tools must not make the final decision, and NSW guidance says AI should not filter out applications. A private employer should have a person review the evidence before any candidate is rejected.
Do employers have to tell candidates that AI is being used?
Australian Public Service agencies must tell candidates in the job advertisement. For private employers, the privacy regulator says businesses should update privacy policies and notices with clear information about their use of AI. Telling candidates up front, in the advertisement and before any recorded assessment, is the safest practice, and it builds trust.
How do you reduce AI bias in recruitment software?
Ask the vendor how the tool was tested, including with Australian candidates from diverse groups. Then keep people reviewing its output, avoid rejecting anyone on a single AI assessment, and audit outcomes regularly for patterns of exclusion. Do not rely on AI to remove bias, because it can reproduce the bias in its training data.
Which AI tool is best for recruitment?
No single tool is best for every employer. Choose on fit with your hiring volume, your roles, and your existing HR and payroll systems. Then test each vendor on four points: who makes the decision, how bias was tested, what candidates are told, and where candidate information is stored.
About the author
Graham Martin is Co-founder of Worknice, an Australian HRIS built for mid-to-large organisations. He has spent more than a decade working with Australian People and Culture teams on HR systems and payroll integration.
This article is general information only and is not legal advice. For advice on your specific obligations, speak to a qualified professional.
Related reading
- AI performance reviews: what helps and what to avoid: the same drafting-versus-deciding split, applied to review cycles and ratings.
- Best AI HR software in Australia: how six HR platforms handle AI across the rest of the employee record.
- The 7 best ATS platforms for Australian businesses: a comparison of applicant tracking systems for mid-market and large employers.
Sources
- Australian Public Service Commission. “AI in recruitment.” 20 April 2026. https://www.apsc.gov.au/working-aps/hr-practitioners/recruitment-aps/ai-recruitment
- Australian Public Service Commission. “Principles for agency use of AI in Recruitment.” April 2026. https://www.apsc.gov.au/sites/default/files/2026-04/Principles%20for%20agency%20use%20of%20AI%20in%20recruitment_accessible.pdf
- NSW Office of the Public Service Commissioner. “Commissioner’s Spotlight: Use of AI in recruitment.” July 2025. https://www.nsw.gov.au/departments-and-agencies/premiers-department/office-of-public-service-commissioner/resources/commissioners-spotlight-on-use-of-ai-recruitment
- Merit Protection Commissioner. “Guidance material for using AI-assisted recruitment tools.” 29 November 2022. https://www.mpc.gov.au/resources/guidance/myth-busting-ai-assisted-and-automated-recruitment-tools
- Office of the Australian Information Commissioner. “New resources on transparency for use of AI and automated decision-making.” 30 September 2026. https://www.oaic.gov.au/news/media-centre/new-resources-on-transparency-for-use-of-ai-and-automated-decision-making
- OAIC. “Guidance on privacy and the use of commercially available AI products.” 21 October 2024, updated 17 January 2025. https://www.oaic.gov.au/privacy/privacy-guidance-for-organisations-and-government-agencies/guidance-on-privacy-and-the-use-of-commercially-available-ai-products
- OAIC. “Employee records exemption.” December 2015. https://www.oaic.gov.au/privacy/privacy-guidance-for-organisations-and-government-agencies/organisations/employee-records-exemption
- Australian Human Rights Commission. “A quick guide to Australian discrimination laws.” https://humanrights.gov.au/resource-hub/guides-for-impacted-individuals/a-quick-guide-to-australian-discrimination-laws
- Diversity Council Australia. “Stage 3: Inclusive AI at Work in Recruitment.” 13 November 2023. https://www.dca.org.au/research/inclusive-ai-recruitment-employer-guidelines