Process-based assessment for AI in Australian higher education

Throughout May, Turnitin brought together higher education leaders from the University of New South Wales, the University of Queensland, RMIT, and other leading institutions to discuss responses to AI in Australian education.

I was expecting the discussion to focus on technology, teaching, and learning. But much deeper themes emerged. Namely, the value of human capabilities in an increasingly machine-driven age, and the importance of struggle in student development.

How can institutions protect human capabilities in the age of AI?

At the heart of AI in education discussions is the need to protect and develop human capabilities. When students use AI to complete assignments, they bypass the struggle they need to truly understand a subject.

In Sydney, Associate Professor Lyn Gribble from UNSW Business School captured the challenge: “There has to be some learning friction. There has to be a moment of what we might even call learning struggle—that moment when you go, ‘Ah, I get it now.’”

Without friction, without pushing to break through learning barriers, students’ long-term understanding and critical thinking skills are at risk. As AI removes that friction, universities are challenged to redesign learning experiences that preserve the cognitive work that matters most.

Around the table, participants discussed how AI isn’t replacing human skills, but changing where humans add the most value. This is critical for designing courses that deliver graduate skills of the future. As content generation becomes automated, interrogating, contextualising, and challenging AI outputs are emerging as key capabilities.

Distinguished Professor Karin Verspoor, Executive Dean of Computing Studies at RMIT, argued that the real challenge isn't teaching students how to use AI. It's teaching them how to think about what AI produces. The graduates who thrive won't necessarily be the ones who know how to prompt AI best. They'll be the ones who know when AI is wrong, biased, or insufficient.

How can leaders embed learning assurance into assessment?

When AI can quickly produce polished outputs, the assessment artefact itself can no longer be the measure of excellence, argued University of Queensland’s Associate Professor Rachel Fitzgerald. Instead, universities must place greater value on visible thinking and human judgment.

This is prompting institutions to rethink assessment design. Participants agreed that Assurance of Learning cannot be layered onto courses at the end of the process; it must be embedded into curriculum design from the outset.

RMIT helpfully shared their Learning Assurance Architecture as an example, showing how they design learning experiences that make core capabilities observable.

“You're not checking the final output that the student produced, but actually the process that they followed to get to it,” explained Professor Michael Cowling, before Dr Alvedi and Ishpal Sandhu, RMIT’s Principal Advisor for Learning and Teaching, went into the details of the framework.

RMIT’s approach enables institutions to map capabilities, scaffold development, and gather trustworthy evidence of learning through:

• Process-based assessment that moves away from summative high-stakes assessment to continuous evaluation of learning
• Dual-purpose rubrics that assess both knowledge and wider graduate skill development
• Scaffolded capability mapping to track each student’s development journey

Alongside this shift in assessment methods, there was also discussion of how institutional AI policies need to evolve in parallel – from prohibiting student AI use and detecting breaches, to proactively guiding students on what constitutes appropriate use, and correcting them when they overstep agreed guardrails.

My top takeaway

Students are already widely adopting AI, and both institutional policy and pedagogical practice need to reflect this reality.

The priority for universities is to design learning experiences that preserve the critical thinking, judgement, and disciplinary understanding that remain uniquely human.

Process-based assessment and visibility into student learning practices are emerging as key ways to do this. To support this, Turnitin has worked with global educators to develop Turnitin Clarity, which provides educators with visibility into student composition practices and development over time without adding additional burden to faculty.

Discover how Turnitin Clarity can help your institution secure learning outcomes.