Australian higher education has moved into a new phase in its response to generative artificial intelligence (AI). The sector has progressed from initial reactions to implementing considered strategies, and a new resource released this week by the Tertiary Education Quality and Standards Agency (TEQSA) provides a practical guide for the paths ahead.
Our previous resource, Assessment Reform for the Age of Artificial Intelligence, was designed as a compass to provide principles. This new resource, Enacting Assessment Reform in a Time of Artificial Intelligence, works towards a map. It outlines the distinct strategic pathways institutions are taking to ensure the integrity of their qualifications in a world where AI is ubiquitous.
It has been nearly three years since the release of ChatGPT. It is now widely acknowledged that student use of generative AI is widespread, and detecting this use is not viable (check out Mark Bassett’s thorough work if you are still in any doubt). This reality presents a direct challenge to assuring that students have met learning outcomes, a core requirement of the Higher Education Standards Framework.
Rather than investing in a technological arms race of detection, the focus must shift to systemic assessment redesign. The new resource details three dominant strategies that have emerged from institutional action plans and sector-wide consultation. There is no single solution to this wicked problem, but we attempted to offer guidance for leaders to help them understand the emerging implications of these strategies.
The three pathways are:
- Program-wide reform. This is a comprehensive approach that redesigns assessment as a connected system across an entire degree. It focuses on the developmental journey of a student, with multiple, secure assessment points that build towards program-level outcomes. While complex to implement, it offers the most robust and holistic assurance of learning.
- Unit-level assurance. This more direct approach embeds at least one secure assessment task, such as an interactive oral task or in-class demonstration (yes, this could also be an exam), within every unit of study. It can be implemented quickly within existing academic structures and provides immediate, tangible evidence of integrity at the unit level. Still, it risks creating a fragmented student experience and a high administrative workload.
- A hybrid model. This pathway offers a strategic blend, applying program-wide principles to critical parts of a degree while retaining unit-level assurance in others. This flexible approach is well-suited to large, complex institutions, allowing for staged reform and adaptation to different disciplinary needs.
What makes this work distinctive on a global scale is not just the content of the map, but how it was drawn. The resource is the product of a deeply collaborative, sector-wide effort, co-authored by dozens of experts from across Australian higher education and facilitated by TEQSA’s Higher Education Integrity Unit. The strength of this national, partnership-based model is that it drew on the expertise of researchers, educators, institutional leaders, educational designers, and academic integrity professionals to co-design guidance for a shared challenge.
However, as we note in the resource, there are ongoing challenges, particularly the vulnerability of fully online and asynchronous assessment and the risk of reverting to examination-based formats that may assure learning but fail to promote it in a way that makes it stick.
This map provides the sector with a shared language and a view of the terrain. We aimed to enable institutions and learning and teaching leaders to navigate the complexities of AI with a collective understanding of the pathways emerging. The task now is to find ways to continue collaborating on mapping the paths forward as TEQSA transitions to a regulatory approach for managing the risks of AI to academic integrity in 2026.
Professor Jason Lodge is Director of the Learning, Instruction and Technology Lab in the School of Education at UQ.