When AI Detection Backfires — Australian Catholic University Drops Turnitin's AI-Cheating Tool

In 2023, plagiarism-detection vendor Turnitin added an AI-writing detector to its toolkit, and Australian universities — including Australian Catholic University (ACU) — began using it to flag student work suspected of being AI-generated. At least a dozen of Australia's 43 universities, and by some estimates around 30, adopted similar AI-detection tools.

In 2024, ACU logged close to 6,000 academic-misconduct referrals across its nine campuses, roughly 90% of them related to AI use — the university's highest number of cases on record. Around a quarter of all referrals were dismissed once investigated, and cases based solely on the Turnitin AI score were dismissed automatically, consistent with Turnitin's own guidance that its "AI report should not be used as the sole basis for adverse actions against a student." Students described transcripts marked "results withheld" for months, and one nursing student said a wrongful accusation cost her a graduate job offer.

In March 2025, ACU abandoned the Turnitin AI-detection tool after concluding it was not reliable enough to support disciplinary decisions. Deputy Vice-Chancellor Tania Broadley confirmed that "around one-quarter of all referrals were dismissed following investigation" and acknowledged that some investigations "were not always as timely as they should have been," while pointing to "significant improvements" over the following year. ACU was one of several universities worldwide — alongside the University of Waterloo (September 2025), Curtin University (January 2026), UCLA and UC San Diego (2024) — to discontinue AI-detection tools after similar reliability and fairness concerns, including independent findings that detectors misclassify non-native-English writing as AI-generated at far higher rates than native-English writing.

This is a cautionary case, not a model to replicate: it shows that deploying an unvalidated AI-detection tool at scale, without robust human review, produced a high rate of unsound misconduct findings and real harm to students, including delayed graduations and lost job offers — evidence of what to avoid when introducing AI into academic-integrity processes.

Read the full analysis: https://www.abc.net.au/news/2025-10-09/artificial-intelligence-cheating-australian-catholic-university/105863524