How AI is being used to stop Deakin University students cheating

Sophie is in her third year at Deakin, she’s never once sat an exam in person. Here’s what AI has to do with it.

Sophie Adams is in her third year of a double nursing and psychology degree at Deakin and has never done an exam in person. AI tech tracks her eyes to prevent cheating, and monitors watch her through her camera. Picture: Alan Barber

Sophie Adams is in her third year of a double nursing and psychology degree at Deakin and has never been to an exam hall.

Instead, she sits at home in Ocean Grove while Artificial Intelligence (AI) technology tracks her eyes to prevent cheating, and exam monitors watch her through her computer’s camera.

“You log in show your student ID to the computer camera and have a photo taken,” she said.

“The program then tells you can’t have anything else open on your computer, and you can’t progress to the exam until you’ve closed all tabs and have shared your screen.

AI tech tracks Sophie Adams’ eyes to prevent cheating, and monitors watch her through her camera. Picture: Alan Barber

“You’re locked into the screen and it’s recording your screen.

“You’re being monitored by a supervisors through the computer’s camera.”

Ms Adams said students had been told the AI system would monitored eye movement to prevent cheating.

“If you’re looking at a phone or another device for too long it watches your eye movement and it will alert whoever’s on the other side monitoring cameras,” Ms Adams said.

Ms Adams said more than 10 of her online closed-book exams had been conducted like that – making up her marks alongside placements and in-person nursing practicals.

“I sat exams in person at school, but at uni I’ve only ever done exams online,” Ms Adams said.

“I do find the whole progression of AI very strange but on a personal level I’d rather sit at home where I’m comfortable.”

Ms Adams said it had been incredibly helpful last year, when she had to return home to her remote Tasmanian hometown for surgery.

“I’ve completed exams from Tassie and other states around Australia,” she said.

Deakin Professor Phill Dawson, the co-director of Deakin’s Centre for Research in Assessment and Digital Learning, said AI enabled those in regional areas or those raising children to sit exams.

“I’ve spoken with students who have told me they wouldn’t have been able to do exams otherwise,” he said.

He said there were several different models of remote exams universities could choose from, and AI eye monitoring was a “common feature” of remote exams.

“It lets the test supervisor know if the test taker has been looking at unauthorised materials.”

He said other models included supervisors watching in real time, or screens being recorded.

Mr Dawson said while they weren’t cheat-proof, there was good evidence that supervised online exams detected cheating.

“But we haven’t had any major breaching,” he said.

Deakin University Deputy Vice-Chancellor Academic Professor Jessica Vanderlelie said about one-third of the university’s students were enrolled in a completely online course.

She said Deakin was taking a proactive approach to embedding contemporary assessments and managing academic integrity in a time where AI is ubiquitous.

“We are drawing on high quality technology to ensure we can appropriately supervise a student while undertaking the exam and be confident in the security of the assessment,” Ms Vanderlelie said.

She said students were supported to set-up their devices for the examination and staff supervised or reviewed when needed.

Originally published as How AI is being used to stop Deakin University students cheating