From AI-hesitance to dialogue: A case study in gen AI-supported assessments
Associate Professor Liz Dale
MEDI993
Faculty of Science, Medicine and Health (SMAH)
Associate Professor Liz Dale designed a series of assessments in MEDI933, a capstone subject in the Master of Indigenous Health, to scaffold AI literacy, while creating a safe space for students to express their hesitancy or concerns about using generative AI (gen AI). Liz recognised that some students would be reluctant to use this technology in academic contexts, whether due to fears about academic integrity or concerns about Eurocentric bias in gen AI outputs. Liz’s knowledge of the impact of gen AI on industry and her commitment to ensuring that her students are work and world-ready motivated her to redesign her assessments in her capstone subject. This redesign intentionally scaffolded reflective practice and created the psychosocial safety needed for her students to “participate ethically and actively in a society where AI is ubiquitous” (Lodge et al., 2023).
Creating cultural and psychosocial safety from the beginning
Liz’s integration of gen AI in her assessment design aimed to develop students’ gen AI literacy in a way that was ethical, transparent and grounded in Indigenous health contexts. The first assessment task in MEDI933 was an ungraded formative task consisting of a Multiple-Choice Quiz (MCQ) on Rapid Literature Reviews and two short answer questions which asked students to reflect on:
- key concepts they have learned in their degree and three things they would like to achieve in the MEDI933 subject
- gen AI and how they have used it, how useful and/or reliable the outputs were and on the biases or limitations students encountered when using gen AI. Alternatively, if students did not use gen AI they were asked to reflect on their reasons for not having used it considering the role it plays in their future career.
We hear Liz Dale from the Faculty of Science, Medicine and Health share findings about students’ gen AI use and their perceptions of gen AI.
Clear gen AI expectations in assessment design to support critical engagement and reduce cognitive load
For their final assessment, students were required to write a Rapid Literature Review as if they were submitting it to a journal. After carefully exploring gen AI herself and considering how this technology could enhance her students’ learning in this capstone subject, Liz drew on Professor Danny Liu’s Menus, not traffic lights: A different way to think about AI and assessments to encourage students to use gen AI for purposes such as - summarising literature, improving clarity or shortening text.
Liz scaffolded gen AI engagement across the subject, moving from open reflection to structured, accountable use. Reflective prompts were provided in both assessment outlines to guide students’ reflections on how they interacted with gen AI tools and their assessment of the biases and limitations of gen AI outputs. “By embedding clear, actionable expectations for (gen) AI use into the tasks themselves” (Corbin et al., 2025, p. 714) Liz reduced “gen AI ambiguity” and the “cognitive load and anxiety” students feel when navigating AI use in assessment. Furthermore, the reflective prompts Liz provided her students with, put cultural safety and ethical responsibility in Indigenous health aspects at the forefront, supporting the priority area of Cultural Safety within the Healing and Recognition Track plan (2024). This approach reinforced that gen AI could support technical aspects of academic work, such as structure and clarity, but should not be relied upon for culturally sensitive knowledge or insight.
The two short answer questions in the first assignment that asked students to reflect on their use of gen AI or their reasoning for not using gen AI, provided Liz with enough context to identify if there were students in her subject that objected to using gen AI. This context was valuable as it informed her decision to include a gen AI statement criterion in the rubric for the final assessment (Figure 1) and it also informed her decision to redistribute the weighting of this criterion for students who identified themselves as non gen AI users. The gen AI statement criterion was also linked to specific Subject Learning Outcomes, further solidifying the relevance of its integration in this assessment.
| Criteria | Exceeds Exceptional (10.0) | Expectations (7.5) | Meets Expectations (5.0) | Emerging (2.5) | Not Observed (0) |
| GenAI statement (5%) (SLO1, SLO3, SLO4, SLO5) |
Integrates the ethical use of GenAI tools clearly and in detail, providing thorough information about how they were used in the literature review and/or report writing process. Provides a detailed discussion on GenAI use, reflecting on its significant role in the process. Consistently acknowledges and appropriately references all GenAI use. |
Integrates the ethical use of GenAI tools with exceptional clarity and depth, providing comprehensive and insightful information about how they were used in the literature review and/or report writing process. Provides a critical, in-depth discussion on GenAI use, reflecting on its substantial role in the process. Acknowledges and appropriately references all GenAI use. |
Integrates the ethical use of GenAI tools, providing adequate information about how they were used in the literature review and/or report writing process. Provides a discussion on GenAI use, reflecting on its role in the process. Acknowledges and appropriately references all GenAI use. |
Integrates the ethical use of GenAI tools, but the information about how they were used in the literature review and/or report writing process may be limited or unclear. Provides some discussion on AI use, but the reflection on its role in the process may be superficial. Partially acknowledges and references GenAI use. |
Insufficient information about GenAI use. Provides little to no discussion on GenAI use, with weak or no reflection on its role in the process. Does not acknowledge or reference GenAI use. |
(Figure 1)
Liz shares how her students used gen AI in their final assessment and how they addressed the biases they identified in gen AI outputs.
Importantly all of these uses aligned with how we encouraged them to do so as we outlined in the assessment guidelines.Students identified significant biases in Copilot and ChatGPT, including a reliance on western research paradigms, the presence of colonial assumptions and deficit-based narratives about Indigenous people.
Once identified, they responded by cross-checking outputs with Indigenous authored resources, avoided gen AI for culturally sensitive content, and stated within the questions asking about limitations that gen AI was best used for structure, grammar and clarity, not generating insights into Aboriginal health.
Impacts and reflections
Across both assessments, Liz gathered rich evidence of how students engaged with gen AI (Lodge et al., 2023) and how their confidence and critical capacity developed over time. Students drew on their cultural knowledge, lived experience and professional roles to evaluate gen AI outputs, assessing reliability not only through academic standards but also through ethical and cultural lenses.
Reflections became increasingly detailed across the subject, with students demonstrating more deliberate interrogation of bias, reliability and cultural appropriateness. This progression suggests that course-specific‑ gen AI resources, when designed as supportive rather than punitive, can help students build confidence and develop ethically and culturally meaningful gen AI literacy.
Liz describes how her students’ gen AI literacy improved, as their confidence grew and their ability to assess gen AI outputs was enhanced.
Across all four assessment cover sheets, I observed that students’ reflections became increasingly detailed, showed more active and deliberate interrogation of bias, reliability and cultural appropriateness. This suggested to me that course specific gen AI resources can offer our students a safe way to engage with gen AI, build confidence and develop ethically and culturally meaningful gen AI literacy.
Liz’s experience highlights the value of designing gen AI assessments that prioritise clarity, cultural safety and student trust. Creating space for hesitation, alongside explicit guidance on acceptable use, supports students to engage thoughtfully rather than defensively with gen AI. Reflective artefacts that focus on learning and ethical judgement, rather than surveillance, are effective in building confidence.
Ever the reflective practitioner, Liz continues to build on student feedback and reflections, and is now refining her approach to further support ethical engagement, preservation of academic skills and cultural safety. A key focus for Liz is ensuring that the gen AI resources she provides her students with remain supportive and manageable and encourage ongoing development rather than compliance.
We hear about the modifications Liz plans on implementing as a result of student reflections and the growing need for students to possess gen AI literacy.
One challenge I foresee is to create a resource that is user friendly where students feel safe and supported to continue developing gen AI literacy and that is not arduous or adds extra work to their study load. So step one will be to reframe the cover sheet as a reflection, moving away from the term “statement” so that students do not feel interrogated, but rather safe and supported in their gen AI experiences and explorations. I will then create specific sections for each skillset containing prompt questions to obtain examples of inputs, outputs and interactions with and that evoke more critical thinking and cultural reflexivity, as this skill is particularly relevant for an Indigenous education context.
Assistant
