The short answer
Using an AI detector to screen job applications or staff writing is risky because the tools produce false positives, and research shows they are more likely to flag non-native English writers. Rejecting or disciplining someone on a detector score alone can be unfair and could expose an employer to discrimination or procedural complaints. If authenticity matters, check it with a work sample, a short interview or a conversation, and treat any detector score as, at most, a reason to ask questions.
- Detector scores are probabilities, and false positives are well documented.
- Non-native English writers are flagged more often in published research.
- Acting on a score alone risks unfair treatment and legal exposure.
- Verify skills with live tasks and conversation instead.
Why employers are tempted
AI tools now write cover letters and application answers in seconds, and recruiters see a rise in near-identical, polished applications. A detector that promises to catch them looks like a simple filter. The trouble is that it filters on writing style, which is only loosely connected to whether the candidate can do the job or wrote the words themselves. Volume is the real driver. When a role attracts hundreds of applicants, any shortcut feels attractive, which is exactly when a flawed one does the most damage.
The false-positive problem
Detectors estimate probability, and no vendor claims perfection. Turnitin, for example, acknowledges that its scores should not be the sole basis for action, and OpenAI withdrew its own detector for poor accuracy. In a hiring pile, even a small error rate means real, honest candidates are marked down for no reason. If one in a hundred honest applicants is wrongly flagged, a pile of five hundred applications means about five people lose out on a mistake they never made. Sources: Turnitin: AI writing detection update from the chief product officer, The Register: OpenAI quietly shuts down its AI classifier
Bias against some candidates
A Stanford-led study found that detectors misclassified over half of essays by non-native English writers as AI-generated. For an Australian employer, a tool that disproportionately flags migrants, or people who write plainly, can turn into an unfair screen on the basis of background. That is a fairness issue and potentially a discrimination one. Candidates who write in a plain, direct style, who are neurodivergent or who are applying in their second language are among those research suggests may be wrongly flagged. Sources: Liang et al., GPT detectors are biased against non-native English writers
Legal and regulatory context
Australia has no law written specifically for AI recruitment tools, but existing duties still apply, including anti-discrimination law, fair process and privacy. From 10 December 2026, organisations covered by the Privacy Act must describe in their privacy policy where a computer program makes or informs decisions that significantly affect individuals. The Australian Human Rights Commission publishes an AI and recruitment compliance checklist. Take legal advice for your own situation. Sources: OAIC: transparency for the use of AI and automated decision-making, Australian Human Rights Commission: AI and recruitment compliance checklist
What to do with existing staff
The same applies to staff reports, emails and proposals. If someone uses AI to draft, that may well be fine, and most workplaces now allow it. Write a simple policy about what is acceptable, what must be disclosed and what must stay out of AI tools, and judge staff on the quality and accuracy of the work. Do not discipline on a score.
A fairer way to verify
To check that a candidate can actually do the work, use a short live task, a work sample done in the interview, or a conversation where they explain their thinking. These are harder to fake with AI and fairer than a detector. If a detector flags something, treat it as a reason to ask a friendly question, and keep a record of your reasoning.
Questions people ask
Is it legal to use an AI detector on job applications?
There is no specific ban, but existing discrimination, fairness and privacy duties apply. Get advice before you rely on one.
Should I reject a candidate whose CV scores as AI?
No. A score is not proof, and many honest candidates use AI to polish their applications. Verify skills another way.
Can I ban AI use by staff?
You can set a policy, but a blanket ban is hard to enforce. Clear rules on what is allowed and what data stays out are more practical.
What should a staff AI policy include?
Approved tools, what data must not be entered, when to disclose AI use and who checks the output.
tools.ceo is published by Kamand, which is mentioned on this page. No affiliate links on this page at publication. Product names link to official sites so you can check current pricing. How we work.
