Evening all,
Firstly apologies that there's been a lull in posting this week. My family had my mum's 70th last weekend (I only look and sound grizzled, I'm not actually that old), and then Australia took a public holiday to commemorate the death and sacrifice of hundreds of thousands of its citizens in war by embarking on a traditional orgy of drinking and gambling. Not me, I spent the day in solemn contemplation.
So I know about a dozen posts ago I did promise to talk about cheating, and I've hardly done so at all, just occasionally gesturing in the distance to the massive heard of wildebeast heading toward us. But with the Australian federal regulator for higher education, TEQSA, releasing their contract cheating masterclass, I thought this might be an opportune time. (Pssst, they're free to sign up for anyone with an edu.au email address, so get on it)
*blatant trumpet blowing klaxon*
Some time in 2021 (I think?) my compadre Cath Ellis of UNSW mentioned that she had been approached by TEQSA to develop some resources for training folks in how to detect contract cheating, and she asked me to join her. Being the collegial type I happily agreed. In truth Cath is a force of nature (for good as far as I've ever seen) and I simply hope that I've held up my end as co-lead of that project. But the team pulled together is spectacular. Basically everyone in this country who could bring a valuable individual skillset to the understanding of contract cheating and how to detect it, we asked. It's been a pleasure working with them all, some of whom will hit the road with us later in the year to host workshops around the country. They really are an A-team. We've since discovered new members. While I naturally have to be Mad Murdoch, it's vastly funnier to think of Cath as B.A. Baracus, constantly muttering "I pity the fool" at everyone. Um yes, where was I again? Ah yes, cheating.
Since the release of ChatGPT in November 2022, there's been a lot of people saying things like "contract cheating is dead." I'd quite happily wager that they're wrong. To the contrary, students being able to produce a vaguely plausible essay or chunk of text underestimates both the scope of the contract cheating market, and the reasons why students engage in contract cheating.
The first thing I thought about ChatGPT (besides "Oh this thing is cool and will flatten a lot of current assessment practice) was that the cost model for contract cheating just collapsed. When your major cost was workers in Kenya, India and other places writing for you, and you have no particular need to produce quality writing, generative AI is a godsend. Basically contract cheating providers saw some of their market evaporate, but what remains comes with a far wider profit margin.
And when we think beyond "essay mills" you start to see that these services are alive and well. It's been my experience that aside from essays, other assessments such as quizzes, exams, experimental reports, and even forum posts, are routinely contract cheated by contractors operating from inside our learning management systems. Myself and a dear (and now former) colleague David House proved this at some scale, years ago. I've gotten better since then. I know other unis are finding this occurring in their LMS too.
When we consider why students contract cheat we might also see that AI isn't a cure all for them. Increasingly, I hear from students that the risk of failing is high, and the perceived likelihood of being caught contract cheating or colluding, etc is low. Therefore when they still perceive ChatGPT as having a tail risk, a few hundred bucks and a pass to contract cheat vs ChatGPT, several thousand bucks and telling their parents they failed seems like an easy choice.
So my views about why we should be looking beyond current practice to a better future aren't based on the idea that this new thing (GenAI) has swamped the old thing (contract cheating). No, it has added more strain onto our structures, stacking and complicating the problems we are facing (student disengagement, wage theft, perverse incentives for most people in the system).
Finally, I'd like to offer a word of thanks to Colin Simpson (@gamerlearner on Twitter) for mentioning the Cunning Plan blog post in his Campus Morning Mail review. I thought something had gone horribly wrong with the site when my phone kept telling me people were looking at the site. Hundreds of people. Now over 500 people. I'm astounded, but thanks very much for reading Murdoch's rant and roll.
Anyhow, I think that's probably enough for now,
KM