Future learning with Gen-AI: Reflections on July session with Barbara Oakley and UoA panellists
30 July 2025Image credit: Barbara Oakley
What cognitive science reveals about effective learning has never been more relevant—especially as Gen-AI increasingly shifts how students engage with knowledge.
In July 2025, the Learning Futures Community of Interest hosted a session centred on Barbara Oakley’s recent engagements with the University of Auckland. Oakley opened with a live talk on ‘paradigm cartels’—dominant narratives that shut down scrutiny and slow educational change—drawing on her work with the New Zealand Initiative and recent observations of the education landscape in Aotearoa New Zealand. Three presentations from four UoA colleagues followed, showing how they had taken ideas from Oakley’s April workshops into their own practice.
What follows are concise session summaries, each framed by one reflective question designed to spark thoughts for your teaching practice, assessment design, and evolving use of new learning tools.
Session overview and recordings (staff login required)
Paradigm Cartels
Presenter: Professor Barbara Oakley
Barbara urged teaching staff to challenge ‘paradigm cartels’—ideas so dominant they become hard to question. She pointed to how experts can become inflexible (“science advances one funeral at a time”), how bias persists even when we think we’re objective, and how retrieval practice remains the most robustly evidenced route to lasting learning. She also noted the negative Flynn effect and New Zealand’s declining PISA scores, warning that without change, students will increasingly offload cognitive effort to Gen-AI.
Even as Gen-AI tools become more common, it’s worth asking where retrieval or spaced practice could help students build durable knowledge, and which longstanding teaching habits may need revisiting in light of what we now know about learning.
Oakley’s cognitive-science framing is strongly evidenced. At the same time, many at UoA work within kaupapa Māori, Indigenous, inquiry-based, or relational pedagogies. The question is not either/or, but how to embed well-supported strategies (e.g., retrieval practice) inside pedagogies that honour collective, holistic, and culturally grounded learning.
What if your most familiar teaching practice is the least effective for long-term learning?
Spaced Retrieval Practice & AI-Adapted Assessment
Presenter: Associate Professor Tanya Evans (Faculty of Science)
Tanya redesigned a large mathematics course around spaced retrieval: after-lecture quizzes, collaborative tutorials, take-home tasks, a mid-semester test, and a final exam—each timed to create repeated, expanding opportunities to retrieve and reinforce learning. A trial semester showed fewer fails and more top grades, and the model has since spread to other core maths courses.
To counteract Gen-AI answer production, Evans replaced some traditional assignments with short student-made video explanations, peer-assessed via FeedbackFruits. This shifts credit from the answer to the reasoning and communication, supports integrity, and scales with minimal staff marking.
Could your assessments shift focus from answers to reasoning—and make space for memory that sticks?
Motivation and Mattering for Learning—AI’s Role
Presenter: Dr Marion Blumenstein (Faculty of Science)
Marion reported on a Cogniti Socratic AI tutor pilot in BIOSCI 107 (coordinated by Dr Suzanne Reid). The tool aimed to support metacognition, self-regulation, and persistence, giving students a patient, ‘always-on’ partner to check understanding, generate self-tests, and probe difficult concepts.
Findings: students valued 24/7 access, the ability to ask ‘uncomfortable’ questions, and the option to go beyond course requirements. Risks included performance illusions, overreliance, and potential erosion of collaborative and critical engagement.
How can Gen-AI boost student agency without eroding collaboration or critical thinking?
Neuroscience-Informed and AI-Assisted Learning
Presenters: Sheryll McIntosh & Professor Jason Stephens (Faculty of Arts and Education)
In their course EDUC 121/121G : How People Learn, Sheryll and Jason combine scaffolding, spaced practice, and retrieval with Gen-AI tools to strengthen metacognition and engagement. Students work through collaborative tutorials, active tasks, and reflective exercises, while AI ‘learning companions’ (a Socratic tutor and self-regulation bot) offer immediate, low-stakes feedback.
Over 2,200 AI interactions and survey data suggest students valued these tools for clarifying concepts and building confidence. The course positions AI as supplementary, with clear expectations around ethics, transparency, and assessment integrity.
Where might a simple AI feedback tool help build better study habits—without becoming a crutch?
What tied the sessions together?
Across the four talks:
- Neuroscience-informed strategies reliably lift engagement and understanding
- Gen-AI needs intentional integration—to enhance, not replace, human learning
- Assessment must change to recognise AI’s presence and protect the value of reasoning, explanation, and practice
Many educators remain wary of Gen-AI for good reasons: ethics, equity, integrity, cognitive overreliance, and the velocity of change. Those concerns matter. However, Gen-AI is already shaping how students access, evaluate, and create information. Whether you adopt it or not, understanding how learners are using it is now part of the job. As notions of what it means to learn keeps shifting, so does our responsibility to stay close to the science of learning and keep asking—openly, collectively—what good learning looks like now.
See also
Learning in the age of AI: A teaching event that met the moment
Watch the April 2025 presentations from cognitive science expert Barbara Oakley, with education researcher Michael Johnston.
Retrieval Practice Guides by Cognitive Scientists
Download these resources (from retrievalpractice.org) for free.
Help Students Retain, Organize and Integrate Knowledge
This resource from MIT is a concise Teaching + Learning Lab guide offering hands-on activities and tips to move students beyond surface understanding.