Australian Catholic University (ACU) used AI to flag thousands students of academic misconduct, the majority related to AI use.
Internal documents seen by the ABC show that approximately 6,000 students were accused across ACU’s nine campuses.
The university took months to investigate the issue and several students missed out on graduate opportunities as they waited for an outcome.
Deputy vice-chancellor Tania Broadley said cases were dismissed if the AI detection tool was the only evidence of academic misconduct.
According to the ABC, over a dozen of Australia’s 43 universities use AI to detect academic misconduct.
Some experts say this number is closer to 30.
Australian universities are required to comply with the Tertiary Education Quality and Standards Authority (TEQSA), however can still make their own policies on things like AI use.
These independent policies mean that TEQSA have no data on which universities use AI to detect AI.
According to TEQSA, universities are permitted to use AI detector tools but are recommended to use other means to verify their results.
While Turnitin claims to have a false positive rate of less than one percent, they also recommend educators to use “professional judgment, knowledge of [their] students, and the specific context surrounding the assignment” to measure academic integrity.
According to information on UNSW’s website, assignments are marked using Turnitin’s AI detection however “all decisions are made based on UNSW staff analysis, not the Turnitin report.”