TEQSA Masterclass Workshop on Detecting Contract Cheating

Penny Wheeler, Amanda White, &
Tom Worthington, at the TEQSA Masterclass

Greetings from the TEQSA Masterclass Workshop on Contract cheating detection and deterrence, at the National Museum of Australia in Canberra. This builds on the online Detecting Contract Cheating course from TEQSA. The good point of the course and workshop is that they raise uncomfortable questions for university academics. But I am not so sure the answers presented are the correct ones. Rather than a "Law & Order - Cheating Investigation Unit", I suggest universities should focus on quality assessment to deter cheating, and structure it to prevent cheating students from ever graduating.

I had some concerns about the legalistic approach in the online module leading to discrimination against vulnerable students, and even more so in the face to face workshop. Also TEQSA clearly envisages having centralized cheating investigation units. 

There is also discussion of what are plea bargains, where students receive a lesser penalty, for admitting cheating. However, if there are to be professional investigators, I suggest they will need to be formally qualified. Also those making the decisions will have to have formal training how to make administrative decisions (similar to the Administrative Decision Making course offered by the Australian Public Service Academy). It does not seem reasonable to deny a degree to a student, based on evidence collected and decided by people who do not even have a relevant TAFE certificate.

Having done an online course, and a face to face workshop, with no form of assessment is not sufficient. Government investigators at least have a vocational certificate (such as a Certificate IV in Government Investigations offered by the Canberra Institute of Technology), and something similar is required for university investigators. Similarly, those serving on tribunals have to have some form of training. 

I suggest that universities should look to their vocational education colleagues, who are much more used to dealing with nationally standardized regulations. 

My preference would be to decriminalize cheating, much the way is being done with illicit drug use in Canberra. Students would be given a failing grade where they cheated, and given compulsory training.

A related issue which came up during the workshop is educational institutions complicity in visa fraud. This where someone obtains a student visa, in order to be able to work in Australia. Those who accept such visas are at risk to being exploited by organised crime. It came as a surprise when I discovered the universities were processing visa applications on behalf of government. It would be tempting to return the function to government staff. But those staff may not have the specialized knowledge of education. It may be that educational institutions, particularly smaller ones, should form consortia and pool staff.

The investigation processes illustrated by TEQSA depended on technical means involving Learning Management System logs, document metadata, Internet Protocols and the like. Such procedures will only catch relatively unsophisticated students, and commercial cheating services. It might be useful to build technical means into the university systems to deter such activities. As an example, the tool could ask the student when they login from a different IP address to explain where they are. Similarly when submitting a document, the system could check the metadada on the document for consistency. The contract cheating company could, of course, coach the students to provide plausible answered, but otherwise only the students who can't afford to pay for good cheating will be caught. 

The course and workshop focused on contract cheating. They were designed before generative AI came to public attention. It is unfortunate TEQSA did not anticipate what was coming and so was required to instead react. They are now trying to catch up. An AI topic which has not yet been addresses is AI for teaching and assessment. I took part in a workshop on building an AI tutor chatbot in 2018, which had some interesting implications.