60k uni students at risk of being falsely accused of using AI in assignments, experts warn

Victorian teachers are being warned not to rely on artificial technology detection programs, after a number of uni students were wrongly accused of cheating on assignments.

About 60,000 Australian university students could be falsely accused of using artificial intelligence for assignments each semester if teachers solely relied on AI detection technology, experts have warned.

Academic integrity and computer law researchers are urging educators not to depend on AI detection programs, such as TurnItIn, when determining whether a student’s work is their own.

It follows a rise in Victorians students complaining to their unions and on social media about receiving high AI generated scores for assignments they insist they personally wrote.

University of Melbourne computing and information systems lecturer Dr Shaanan Cohney said despite companies insisting their AI and plagiarism detection programs have a low false positive rate, multiple students are still being wrongly accused of using AI to do their work.

“While a one per cent false positive rate sounds pretty low, when you think about the number of students it could impact, that’s pretty substantial,” he said.

“I teach between 200 and 1000 students, which means every single class we’re going to have multiple students who are falsely accused if we rely on these tools as a single source of truth.”

Deakin University Centre for Research in Assessment and Digital Learning co-director Professor Phillip Dawson said if every Australian university relied on AI detectors, tens of thousands of students would be falsely accused of using the technology for their work.

“(There would be) about 60,000 (false positives) per semester by my calculations,” he said.

Daniel, a second year arts student at Monash University, said he was “wrongly accused” of cheating on an assignment last year.

“I was told by the university that I had come under investigation for misconduct due to cheating on my assignment, but I didn’t use AI or any cheating software for it,” he said.

“It was a very stressful situation because I had done nothing wrong and that was eventually proven.”

Monash Student Association president Chloe Ward said lengthy and confusing artificial intelligence policies were also leaving students vulnerable to AI misuse in assignments.

“Students are facing an increased number of academic misconduct charges due to the murky

rules around AI use in assignments, research and assessments,” she said.

“Some units allow for students to use Generative AI for idea generation but not their final work. So where is the line?”

University students claim they are being accused of using artificial intelligence for assignments.

National Union of Students president Ngaire Bogemann said the rise in students being falsely accused of using AI was of particular concern for international students.

“A recent report from Stanford University has revealed AI detection tools are more likely to identify text written by non-native English speakers as AI-generated,” she said.

As for how students can avoid being falsely accused of using AI, Dr Cohney said students should “not use AI for any part of their assignment process”.

Prof Dawson said students should also keep version history switched on when they are working on their assignments and could consider emailing themselves copies of their work in progress.

A University of Melbourne spokesman said the institution “requires staff to consider additional evidence before making an allegation of academic misconduct”.

A Deakin University spokeswoman said students are also asked to use AI tools responsibly, while Australian Catholic University provost and deputy vice-chancellor (academic) Professor Julie Cogin said the university has strengthened its policies to recognise new forms of cheating.

A TurnItIn spokeswoman said the company regularly tests its AI writing indicator against 800,000 academic papers to maintain a false positive rate of less than 1 per cent.

“Our AI writing indicator is designed to serve as a data point and provide information that educators can use to drive productive discussions and student engagement,” she said.

“Educators are best positioned to make decisions around academic integrity. We encourage ongoing independent research and invite universities to run their own tests.”