Universities should learn lessons from the tax office to improve security of assessment, refocus their attention on improved assessment of student work and start embracing AI in courses immediately, according to expert advice this week.
In a wide ranging presentation to a Victoria University symposium on academic integrity and assessment security, University of Sydney Academic Director Professor Cath Ellis said that universities made a compelling pitch for universities to rethink the approach to teaching and assessment.
Professor Ellis said universities couldn’t wait a couple of years to work out how to incorporate AI tools into teaching and assessment, as industry were using these tools right now.
“Otherwise we are going to produce unemployable graduates,” Professor Ellis said.
“In a world where Chat GPT exists, what is the work (that is relevant for students to do)?
“We need to shift our energy from worrying whether cheating has occurred to worrying whether learning has occurred.”
Professor Ellis said the higher education sector should start to learn from other sectors in developing new approaches to policing cheating and ensuring security and validity of assessments.
Other regulatory organisations had developed sophisticated models focused on policing the individuals who represented the highest reputational and public risk, while providing tools and processes to modify behaviours into an acceptable range for lower risk individuals.
Higher education organisations needed to spend more time investing in enforced self regulation – organisational processes and practices which would encourage students to do the right thing. This approach was implemented successfully at airports, where passengers were told not to take knives or other objects onto planes, and most did, but the presence of security scanners after check-in enforced acceptable behaviours. This is clearly a more effective approach than allowing all passengers through and then trying to catch offenders.
Professor Ellis said universities needed to re-centre their approach to teaching and assessment and instead of obsessively focusing on how AI can increase the risk of cheating, focus also on how they can now prove that learning has occurred.
“We need to empty the value of cheating from our courses and get to a point where cheating is pointless,” Professor Ellis said.
“No (learning) task can ever be completely secured.”
New approaches to teaching delivery paired with alternative approaches to assessment were critical to the future of teaching and assessment, she said, providing an example of improved outcomes through oral assessments.
