Recent reports from Australia and overseas show that technology-enabled cheating is moving well beyond students copying from the internet or using ChatGPT to draft an essay. At the University of Sydney, Honi Soit (the University of Sydney student magazine) reported that ten subjects had been affected by leaked exam materials, with the issue first surfacing in ‘Introductory Microeconomics’ and then extending to mathematics, data science, business, biology, computer science, chemistry, engineering, economics and information systems. Students reported that exam material appeared on Chinese social media platforms and that part of an exam may have been filmed or live-streamed. For VET educators, the key issue is obviously not the university setting where these problems occured, but the fact that assessments can now be captured, shared and even monetised almost instantly by students.
A related concern is the continuing evolution of organised cheating services that now combine social media recruitment, covert devices and real-time assistance. Farrago Magazine (the University of Melbourne student magazine) reported allegations that contract cheating syndicates were active around University of Melbourne exams, including claims that students were being supplied with hidden cameras and covert earpieces so that exam questions could be transmitted to others and answers could be fed back to students in real time. While the claims had not been independently verified, they align with broader warnings from TEQSA that commercial academic cheating services are using direct promotional tactics online and in class messaging groups. TEQSA has also warned that students who engage with these services can be exposed to blackmail, identity theft and cyber security risks.
Generative AI is also changing more ordinary forms of assessment misconduct. ABC News reported that Mazenod College in Melbourne identified students who had used AI tools in a Year 12 oral English assessment. With a number of assessments in VET requiring students to prepare oral presentations, case study responses, portfolios and role-plays, it is important that educators and crucially assessors understand that these are also now tasks where AI can be used to produce fluent, plausible work that may not represent the student’s own skills.
Recent international examples also show that the risks to academic integrity are not limited to generative AI text-based tools. In India, Reuters reported that the government temporarily blocked the Telegram platform, after allegations that it had been used for organised cheating in relation to a national medical entrance test, and the Wall Street Journal subsequently reported extraordinary security measures brought in by the Indian government involving helicopters, air force jets and paramilitary escorts to protect test papers.
In Thailand, the South China Morning Post reported allegations that officials altered computerised civil service exam scores in exchange for bribes. The Nikkei Asia was reporting that Chinese authorities are now actively checking if students are wearing AI-enabled smart glasses to take assessments, as well as using AI in exam halls to monitor suspicious student behaviour “whispering, looking around, copying from a neighbour, answering (early) or continuing to write after times is called.”
Together these examples show that assessment integrity now needs to go beyond plagiarism checking and include device management, platform monitoring, secure handling of assessment materials, identity verification, and protection of back-end assessment systems.
For the VET sector, the practical response is to strengthen assessment design rather than rely only on detection. With ASQA shortly due to release new principles for the responsible use of AI, RTOs need to be thinking about giving students clear guidance on permitted and prohibited AI use, redesigning potentially ‘vulnerable’ assessment tasks (within the parameters allowed by the relevant Training Package), build in observed practical performance, use oral questioning to verify understanding, and ensure staff know how to identify suspicious patterns.
The aim for the sector needs to be making sure every assessment judgement is based on genuine evidence that the student can demonstrate the required skills and knowledge themselves.
