AI-detection software fails too often; schools and unis are ditching it
AI-detection tools are so unreliable that NSW stopped schools from relying on them to decide if students cheated. Universities are phasing them out.
Schools and universities have used AI-detection software to catch students using ChatGPT. Now NSW has stopped schools from relying on these tools.
Research showed they are not reliable enough. Even when companies claim false-positive rates below 1%, that still means over 100,000 wrong accusations per year across Australian universities alone.
Australian Catholic University flagged nearly 6,000 students for AI cheating in 2024. The university later dismissed cases where the evidence came only from faulty software.
Students wrongly accused of cheating report serious stress and anxiety. Some universities including Curtin and Queensland have stopped using the tools altogether.
Others including Melbourne, Sydney and UNSW still use them but only to flag cases for human review.
Why it mattersStudents can be wrongly accused and harmed by software that is not accurate enough to decide guilt. Better judgment calls are needed on academic honesty.
AustraliaAustralian students at universities using AI-detection tools face a small but real risk of false accusation; NSW schools now have explicit protection against this.
✓ Claims checked against the source and corrected before publish. checked 10 h ago
