So, as educators, we face a conundrum: how do we help students develop evaluative judgement while they are still building the disciplinary knowledge needed to make those judgements well?
The following examples show some ways of how this can look in practice. Each case creates a structured moment where students do more than use Gen AI: they pause, compare, question, explain, and make judgements about the role AI should play in their learning. These moments help students recognise and articulate how learning occurs, making their thinking and decision-making processes explicit (Lodge et al., 2026).
In a STEM Machine Learning course, educators Dr Zhendong Huang and Dr Shuwen Hu encourage students to use an AI tool for a project assessment task, but they must also describe how they used it. Students must explain which tool they used, what they used it for, how they checked and verified outputs, and what improvements they made with AI assistance. The task requires students to disclose, explain, validate, and reflect on their use of AI.
Another example comes from Jo Bradley at RMIT UP Foundation Studies in STEM. Students first created their own mind maps and recorded short explanations of the concepts and connections they had identified. These artefacts were then used to prompt Google Gemini to generate a consolidated mind map. Students critically reviewed the AI-generated version for accuracy, completeness, misconceptions, and missing information. The result was a student-verified study worksheet that supported revision while developing students’ capacity to evaluate AI-generated content.
These examples also reinforce Lodge et al.’s point that students who learn and think ethically with Gen AI need more than tool skills. They draw on regulatory skills, ethical judgement, and disciplinary knowledge (2026, p. 10).
Emilio Kardaris, Lecturer in Chiropractic, takes such an approach by using AI role play to help students practise and critique communication before applying it in live settings:
"In a final-year chiropractic course, I use a generative-AI role-play tool to help students practise behaviour-change conversations before performing them live. Students experiment with different approaches and watch the AI scenario shift in response, prompting them to question why, or why not, their communication worked. We then unpack these interactions in class, focusing on the limits of AI simulations, patient nuance, and the importance of clinical reasoning beyond the algorithm. This approach gives students a safe space to think critically about AI’s role in practice while strengthening their ability to guide real patients through meaningful change."
Peter Murphy, Associate Lecturer in Primary Curriculum and Pedagogy in the School of Education, approaches this through collaborative and networked learning:
"In my teaching, I invite students to treat AI as a speculative collaborator within a shared, networked learning environment. Drawing on the principles of Connectivism, we use collaborative digital platforms to make collective thinking visible and open for feedback. For example, I may prompt students with "How might we leverage AI to help manage complex classroom behaviours?" Students share inputs, compare responses, and build on each other's ideas, creating a dynamic record of learning across the cohort. AI becomes one contributor among many, helping the group explore possibilities, challenge assumptions, and reflect critically on emerging knowledge. I believe this approach builds creative confidence and collective critical thinking within an evolving, connected learning community."
Chris Hope, from Business Administration, uses a deliberate pause before AI is introduced into discussion activities:
"In my MBA teaching this year, when I run in-class activities, I ask that AI is not used and students force themselves to first think critically about the question or topic and discuss it with the group. After some time, I encourage the use of AI, but only to challenge and build on their perspectives. In addition, I run a debating activity where students argue a viewpoint against AI. The feedback on this approach has been remarkable, with many comments thanking me for pausing the use of AI in those discussion contexts, while also enjoying the ability to challenge and consider the importance of information literacy when debating against it."
Across these examples, Gen AI is not treated as a shortcut or a replacement for learning. Instead, students are asked to question, test, compare, and use AI with intention, and this is done within the disciplinary context in which they are building knowledge. These activities are designed to help students build the evaluative judgement and adaptive capabilities they need for study, work, and everyday life in an AI-rich world.
These activities, taken in isolation, are not solutions in themselves. However, together they show how educators can create repeated opportunities for students to question, test, compare, and use AI with intention within the disciplinary contexts in which they are building knowledge. We should not expect evaluative judgement to be developed through a single course or task. It is built gradually across a student’s whole program and continues to develop beyond graduation. Like a muscle, it is strengthened each time students are supported to critique, evaluate, question, and think for themselves. In this way, and in line with Lodge et al., these activities help students become lifelong learners who can adapt to evolving technologies, building confidence and capability over time.
