TEQSA backs GenAI learning reform in Australia

Australia’s Tertiary Education Quality and Standards Agency has published a paper on how higher education institutions can assure quality learning in a future shaped by generative AI.

The paper, ‘Assuring quality learning in a GenAI-integrated future: The role of adaptive capabilities’, argues that universities need to rethink how they define, assess and evidence student learning as generative AI becomes embedded in education.

The authors say generative AI and automated decision-making systems challenge traditional approaches to academic integrity and assessment. Rather than focusing only on securing final submissions, institutions should clarify what students need to learn in AI-integrated environments and how that learning can be demonstrated.

The paper identifies adaptive capabilities as central to graduate learning. These include digital literacy, distributed cognition, hybrid metacognition and life-long learning, grounded in disciplinary knowledge and supported by student agency and regulation.

The authors warn that narrow AI literacy may not be enough, as operational skills linked to current tools can quickly become outdated. Adaptive capabilities can help students evaluate new technologies, use AI ethically and continue learning as systems evolve.

The paper also highlights risks linked to generative AI, including overreliance on AI-generated explanations, reduced effortful learning and excessive cognitive offloading. It says higher education should preserve practices that support deeper learning, such as retrieval practice, spaced revision and generating answers before receiving explanations.

Assessment reform is a major theme. The paper calls for greater attention to evidence of learning processes rather than only to final products. Possible approaches include portfolios, learning journey documentation, reflective tasks, trace data and structured self-assessments.

TEQSA says the paper is not prescriptive and does not form part of its formal guidance notes. Instead, it is intended to support institutional thinking about how quality assurance may need to change as generative AI becomes a normal part of higher education.

Why does it matter?

Generative AI is weakening the reliability of product-based assessment, especially when final essays, reports, or problem solutions are produced or heavily shaped by AI tools. TEQSA’s focus on adaptive capabilities points towards a different quality assurance model: one that values student judgement, process evidence, ethical AI use and deep disciplinary understanding. That matters for universities because they will increasingly need to prove not only that students produced work, but that they learned, reasoned and exercised agency while using AI.

Would you like to learn more about AI, tech, and digital diplomacy? If so, ask our Diplo chatbot!