4 Designing with AI: Exploring Inclusive and Patient-Centred Learning in Dentistry

Keywords 

Generative AI, course design, dentistry education, patient-centred learning, Universal Design for Learning (UDL); assessment design

Practitioner Notes

1. Patient-centred case studies, scaffolded with AI support, helped students see connections between professionalism, ethics and their future clinical practice.
2. AI tools streamline course redesign, particularly in structuring LMS content, developing assessment FAQs, and editing personalised feedback.
3. Offering multimodal formats (interactive modules, podcasts with associated course readings, slides, recorded tutorials) aligned with UDL (universal design for learning) and improved accessibility for students with varied needs.
4. Embedding an AI use declaration on assessment coversheets normalised disclosure, raised awareness of responsible practice, and encouraged reflection on integrity.
5. AI-generated images, voices, and slides required careful validation; integrating these AI scaffolds with educator expertise was essential to ensure accuracy, authenticity, and trust in learning materials

Background

What are My Courses About?

I coordinate and teach courses on dental public health and professional practice. These courses incorporate topics and concepts that reside outside the realm of traditional clinical training but remain essential ‘soft’ skills to help students maintain professionalism and cultivate ethically sound practice (Kodali et al., 2024). Professionalism and ethical practice is essential to all health professionals. Dental professionals, along with eleven other health practitioner groups are governed by the Australian Health Practitioner’s Agency (AHPRA) through a shared Code of Conduct. This Code of Conduct sets out the standards of professional conduct the National Boards expect and is used by Boards to oversee health practitioners. All health practitioners have a professional responsibility to be familiar with and to apply the Code of Conduct (AHPRA, 2024; Weerakoon et al., 2025). My courses provide the grounding to help students understand what it means to be a professional, while learning about broader public health concepts and how they relate to everyday clinical practice.

How Do Dental Students Engage with My Course Content?

Students typically experience this course in varying formats that include workshops, short lectures, self-directed reading and online interactive modules on a learning management system known as UQ Extend. In 2024, students experienced a simplified version of this course, as I was still new to teaching it. I largely retained the 2023 structure, though at the request of the now previous Chair of Teaching and Learning, I redesigned assessments to be more resistant to AI use. On reflection, however the 2024 cohort had been challenged to demonstrate greater creativity and engage less with rigid marking rubrics, a useful reminder of how assessment design shapes learning. As the semester and 2024 iteration of the course progressed, it became apparent that many students perceived the broad range of topics as disjointed and disconnected from clinical practice. Another key issue was engagement, workshop attendance was low, with many students reliant on readings alone. Some chose to covertly recorded informal post-workshop conversations, particularly when assessment details were discussed, suggesting clarity around assessment held more value for them than participation in broader learning activities.

In 2023, students accessed course materials through the learning management system, UQ Extend, However, when I assumed responsibility in 2024, I was unable to continue using this platform. The previous coordinator had created the content on an external platform, Genially, a commercial platform for creating interactive and animated visual content such as presentations, infographics and quizzes. While I could share these resources with students, I was unable to adapt the 2023 content or formatting to meet the evolving needs of the course in 2024 and beyond.

GenAI Tools I Engaged with

In the 2025 iteration of my courses, I predominantly used ChatGPT. Although I remain open to experimenting with other platforms to explore what they can offer and how they might support my teaching. I enjoy testing tools, pushing their capabilities, and often use them in unconventional and divergent ways. This exploratory approach led me, somewhat unexpectedly, to create a podcast series for the course, an outcome that illustrates how playful engagement with technology can generate multimodal resources that appeal to different learners. Alongside ChatGPT, I also experimented with Google Notebook LM, Adobe Firefly, OpenAI’s Sora and the AI features built into applications such as Prezi. Each tool opened distinct possibilities. Firefly, for instance, enabled the creation of ethically sourced illustrative visual content that could support patient case studies; Sora suggested future directions for generating short, scenario-based video materials; and Prezi’s AI features streamlined slide development, though always with careful human oversight to ensure accuracy and contextual relevance. These explorations reinforced a broader SoTL insight, that experimenting across platforms can spark creativity and extend inclusive design, but that every AI application requires critical evaluation to ensure it aligns with pedagogical intent and that its use is fit for purpose.

Three Inter-Related Reasons I Chose to Integrate AI into My Courses

I integrated AI into my courses:

1. To demonstrate clinical relevance: embedding public health and professionalism in patient centred contexts to help students gain insight and add relevance towards their clinical practice.
2. To reduce my administrative and design workload: using AI tools as co-creative assistants to streamline routine tasks and free capacity for meaningful teaching work.
3. To diversify learning options: Incorporating multimodal resources aligned with UDL (Ralabate, 2011) principles to support a diverse cohort.

DENT3000 and DENT7221 are the only courses in the dental curriculum that explicitly address the broader worldview concepts of Public Health. Yet, many dental students struggle to appreciate the value of Public Health and associated ‘soft skills’ for their future professional practice (Kodali et al., 2024). This sentiment is captured in student feedback from 2024, prior to AI-integration, as expressed by a student.

Such feedback highlights the challenge of positioning public health as integral to dental education and reinforces why new approaches such as AI-enhanced design are necessary to make the content more engaging and clinically meaningful. I believe that so-called soft skills are undervalued because healthcare programs such as Dentistry often teach clinical skills in silos that reflect clinical and biomedical worldviews (Abdul Kadir & Schütze, 2022). These structured worldviews can fragment student learning and leave little space for the scaffolding of essential human skills within the curriculum (Flaitz et al., 2011).

Public Health subjects require educators to translate broad, abstract and world-view concepts into accessible forms. This is difficulty for mid-level dental students who are only beginning to develop patient care experience and higher-order thinking. Consequently, public health educators need a strong disciplinary foundation and clinical background to contextualise relevance for learners (Abdul Kadir & Schütze, 2022). Against this context, and considering the modest (this is being kind) performance of Student Evaluation of Course and Teachers (SECaTs) results for DENT3000, it became clear I needed to develop, adapt and integrate Public Health principles into clinically relevant scenarios that students could recognise as meaningful for their future practice.

Courses such as Dentistry challenge academics, such as myself, to balance the requirements of maintaining clinical practice with the expectations of teaching, research and service (Nassar et al., 2019). In my case, I own a dental practice and work as a clinician on the Sunshine Coast, Australia while also holding a Teaching and Research Focussed position at the University of Queensland. The mounting academic expectations often stretch my part-time (0.4 full-time equivalent) capacity. In Semester 1, 2025, I was expected to co-teach and coordinate two 2-unit courses, with enrolments increasing from 64 students in DENT3000 (2024) to 107 across DENT3000 and DENT7221. While a small number of guest speakers contributed ad hoc content, the bulk of the design, coordination, teaching and assessment responsibilities rested with me. Meeting these responsibilities allowed me to be open to adopting AI-supported strategies to manage my workload more efficiently and to ensure students continued to experience relevant and engaging learning (Hashem et al., 2024). This became the point at which AI became a necessary teaching partner in my practice.

In 2024, Generative AI was still an emerging and contested technology, and like many institutions, the UQ School of Dentistry initially adopted a cautious stance (Xiao et al., 2023). Senior academics encouraged course coordinators to ‘ban’ AI in student assessments. Yet, the adaptation of AI into assessment practice was, and remains inevitable (Xiao et al., 2023). Dental students are primarily taught and assessed through hands-on activities using typodonts attached to mannequin heads (El-Kishawi et al., 2020). These practical skills, grounded in empirical knowledge, dominate their training, often at the expense of written communication development (Kodali et al., 2024). As a result, many dental students perform poorly in written assessment tasks (Holtzman et al., 2005). Unlike other parallel dental courses at the University of Queensland that rely on short-answer and multiple-choice examination, DENT3000 culminates in language-intensive assessments, such as written and video-based assessments.

Despite the ban, eleven of the sixty-four students in the 2024 DENT3000 cohort submitted written assessments with significant AI-use detected by Turnitin. Although this represents less than 30% of students (as indicated in a survey of AI use) (Gonsalves, 2025). The students were at risk of breaching the University of Queensland’s (then) Academic Integrity policy. This risk was partially mitigated by ongoing controversy over the accuracy of AI detection tools (Malik et al., 2025) but the AI incident nonetheless caused considerable distress for both students and myself as coordinator. I decided to conduct informal interviews with the students. My findings indicated the students who had used AI came from culturally and linguistically diverse backgrounds (CALD) with self-reported lower levels of confidence in written English. This finding reframed the incident for me as it was more than just a case of misconduct, it highlighted that gaps in language support and assessment design can unintentionally disadvantage groups of learners. After extensive consultation with senior academic staff, the students at risk of academic misconduct were asked to re-submit their tasks without the use of AI. For me, this was also a turning point.

Recognising that AI was not going away, I began to self-educate myself on its potential. At the time, AI was very new to the University, and I could not find formal courses or guidelines available through the institution. Thus, my initial self-education process began with books that provided general context and information on AI. Once I understood the basics, I searched for and completed brief online courses that were hosted by a UK-based institution and by academics sharing practice on Twitter/X. By the end of Semester 1, 2024, I was ready to experiment with ChatGPT directly to explore the model’s assistive possibilities for teaching and learning.

During Semester 1, 2024, I also became aware of the number of students with UQ Access Plans, formal arrangements developed through the University’s Disability, Diversity and Inclusion team, to support students with a disability, medical condition, or learning difference. These plans outline reasonable adjustments to help students engage equitably with their studies. Several enrolled students with Access Plans identified learning and cognitive challenges as my course required a high volume of reading. Many of these students reported difficulties balancing reading workload alongside competing demands in other units. The heavy reliance on text-based materials risked creating barriers to participation and underscored the need for more inclusive, multimodal approaches to content delivery.

After much deliberation, I recognised the need to develop learning materials that were both inclusive and clinically relevant, while remaining manageable within my existing academic workload (Ralabate, 2011). My overarching aim was to introduce public health and professionalism content into clinical contexts by applying inclusive learning modalities with AI as my co-creative teaching assistant. For the 2025 delivery of DENT3000 and DENT7221, I created two new UQ Extend sites from scratch, adapted existing teaching materials, and introduced online recorded tutorials which had not previously been part of DENT3000. This process was, in many ways, like developing a brand-new course. I had to ensure it was fit for purpose, ensuring the redesign streamlined processes rather than add further strain. Unlike colleagues who typically plan digital uplifts over 6–12 months, I chose to implement mine in real time. This approach gave me the flexibility to experiment with new platforms, learn tools as I went, and adjust content in response to informal feedback. It also allowed me to gauge whether innovations resonated with the students, and what aspects required further refinement.

One key challenge was explaining to students why we were using a variety of content delivery methods, including recorded lectures, PowerPoint PDFs, UQ Extend modules, and live tutorials. For some, who prefer text-based learning, the UQ Extend modules felt frustrating and time consuming. One student said it was “a waste of time,” citing broken links, constant clicking, and a preference for direct access to PDFs and traditional lecture recordings. For others, however, the interactive modules provided valuable scaffolding that helped them unpack complex content and enhance their learning experience. These contrasting responses underscore a critical SoTL insight: students are not always conscious of how their learning preference shapes their engagement. As educators, we need to make the rationale for multimodal design explicit, not only to diversify delivery but also to meet the varied needs of our cohort (Tai et al., 2021). In this sense, inclusive design is as much about transparency as it is about accessibility.

Methodology

I approached this work as a form of practitioner inquiry using student feedback from surveys and focus groups to improve and guide changes to the course (Keane & Labhrainn, 2005). At the University of Queensland, course coordinators receive formal teaching feedback through the Student Evaluation of Course and Teaching (SECaT) and the Student Evaluation of Tutor surveys (UQ, 2025). These instruments are designed to capture student reflections on both teaching practice and course content, providing an institutional mechanism for evaluating educational experiences. As a course coordinator, my courses and teaching are subject to SECaT surveys each semester. While these surveys offer valuable insights, the anonymous nature of responses can at times feel confronting: the feedback is often a mixture of constructive suggestions and comments that may be perceived as discouraging or overly critical. Even so, engaging with this feedback became a key part of my methodology and informed my experimentation with AI informed course design.

Personally, I found my 2024 SECaT survey outcomes difficult to engage with, because the constructive feedback and comments were interspersed with statements that could be interpreted as derogatory or inappropriate.  Rather than dismissing this experience, I treated it as a point of reflection: how might such feedback be used to generate constructive change. From this starting point, my methodology for this chapter took shape. It serves as a practical outline, a ‘how to’ guide, documenting the digital tools I’ve experimented with, the strategies I adopted and the ways I have shared knowledge more effectively with students. This chapter is grounded in a process of inquiry, using feedback as data to inform cycles of experimentation and reflection.

I began my process by uploading the collated 2024 (Pre-AI integration) SECaT scores and feedback into ChatGPT and asked what can I do to improve this course? The outputs suggested three directions: (1) develop a UQ Extend site, (2) create more interactive online modules and (3) provide clearer guidance on assessment tasks. Consequently, what follows in this chapter is presented as a toolkit for practice. For each digital tool or teaching strategy trialled, I provide context, a sample prompt or workflow, and a reflection on outcomes – highlighting successes and setbacks, as well as opportunities for improvement. Where relevant, I also include links to resources or visual examples to support readers to explore these strategies further, and to maintain the spirit of UDL for you, as the reader.

I designed this chapter with UDL principles in mind with an aim to to include more abilities by offering multiple means of engagement (Ralabate, 2011). Personally, I usually learn by reading and engaging with visual materials, but I recently started listening to audiobooks and podcasts. This shift in how I engage and acquire knowledge prompted me to reconsider how I communicate content to my students, by recognising how diverse formats can create more equitable learning opportunities. Most importantly, I documented everything so that I could explain my process. Consequently, this chapter aims to demystify the integration of AI in course design. If you are curious about engaging with similar tools in your course, some of my ‘experiments’ may serve as a useful starting point or inspire new way of working.

This study received ethics approval from The University of Queensland Human Research Ethics Committee (BEL LNR Panel), project number 2024/HE001346: AI Inquiry-Led Teaching Innovation.

My Experiments and Outcomes with AI

Refining the course based on student feedback.

Context

Engagement with student feedback can be challenging for many academics when constructive feedback is mixed with critical comments. Using AI can create a sense of distance, making it easier to process feedback objectively and translate it into meaningful course improvements (Flodén, 2017).

Objective

To use AI to analyse SECaT survey outcomes for the 2024 DENT3000 course to generate a list of positive attributes and practical suggestions for future course improvement.

Outcomes

The analysis generated a series of recommendations, which I trialled in practice. My responses to each are included here to illustrate how AI-supported feedback informed concrete changes.

1. Assessment Design:

• • Increase word limits for assignments, especially for the grant application (Assignment 3), as many found 500 words insufficient.

Response: The word count for Assignment 3 was increased to 1000 words.

• • Provide clearer rubrics, better exemplars, and more guidance for all assignments.

Response: In addition to introducing assessment tasks in class, I uploaded FAQ documents to the LMS and alerted students to this via announcements.

• • Consider reducing the number of assignments or integrating them into a portfolio format.

Response: This is challenging to do because the Australian Dental Council accreditation requirements restrict such changes.

2. Course Structure and Resources:

• • Consolidate weekly materials into single folders to avoid confusion.

Response: Content was consolidated into weekly folders and a course guide provided at the beginning of semester.

• • Minimize reliance on external links, focusing instead on lecture slides with relevant content.

Response: Some external links referred to sites where students were required to register and enrol in short courses run by organisations such as Quitline to learn about ‘Smoking Cessation’. In 2025, I created new self-directed learning modules in UQ Extend to streamline this.

• • Align self-directed learning, lectures, and workshops to reduce cognitive load.

Response: The 2024 version of DENT3000 followed a flipped classroom where students were required to complete their self-directed learning modules prior to attending workshops. However, many did not engage with the modules until exam preparation time. I revised the 2025 version to align the content with the workshop or tutorial topic for the week.

3. Guest Lectures:

• • Improve the relevance of guest speakers’ content to course objectives.

Response: I expanded course objectives into detailed success criteria, shared with the guest speakers and invited feedback. All contributors and students were supplied with an outline of the learning expectations for individual sessions.

• • Consider live Zoom lectures or pre-recorded sessions for guest lecturers to increase flexibility.

Response: More pre-recorded lectures were uploaded, and tutorials were run online, recorded and transcribed by the AI feature in Zoom.

4. Student Workload:

• • Reduce the volume of weekly readings and ensure they are directly applicable to assessments.

Response: I reduced the number of required readings, and selected essential readings were converted into podcasts to create inclusive learning aligned with learning objectives and success criteria.

5. Technical and Administrative Aspects:

• • Streamline the learning platform to make content navigation easier.

Response: All content was delivered via the UQ standard LMS, Blackboard, with fewer folders and new self-directed modules were created in UQ Extend.

• • Share assignment updates through announcements rather than discussion boards to ensure all students receive them.

Response: All updates were delivered via Blackboard announcements. In addition to individual feedback, students were also given overall feedback as a group.

Use AI to Align Course Objectives and Discern Key Differences Between Co-Taught Courses

Context

In 2024, I coordinated and taught DENT3000. By 2025, I was tasked with coordinating and co-teaching both DENT3000 and DENT7221. DENT3000 is a third-year undergraduate course in the Bachelor of Dental Science (Honours), while DENT7221 is offered in the second year of the Doctor of Dental Medicine (DMD), a 3.5 year full-time graduate entry program for students with prior backgrounds in health, allied health, physical or biological sciences. While both cohorts qualify as general dental practitioners in Australia, the different entry pathways and levels of experience shape course design and complexity for co-teaching. Both courses are 2-Unit courses with a standard full-time student workload set at 8 units.

Objective

To use AI to analyse and compare courses objectives, identifying key differences and similarities between DENT3000 and DENT7221.

Outcomes

This analysis provided a clearer understanding of expectations across the two cohorts. This enabled me to adapt the assessment tasks in DENT3000 to reflect undergraduate context, while maintaining alignment with professional standards. DENT7221 tasks and rubrics, already co-designed with a learning designer served as a useful benchmark. Aligning both courses in this way helped me streamline design decisions and ensure coherence across parallel pathways to the same professional outcome

Communicate Expectations: Enhance and Unpack Learning Objectives

Context

Learning objectives are often written with minimal context and are often open-ended, which can make it difficult for students to understand what is expected of them. The lack detail allows for teaching flexibility, but can leave students uncertain about how to demonstrate achievement. Broad guidelines often make it harder for guest lecturers or tutors to align their contributions with course outcomes.

Objective

To create success criteria that extrapolates a learning objective to provide detailed guidance. An example of a single prompt and output is provided here.

Outcome

Students were provided with a course handbook with detailed success criteria attached to overarching learning objectives. Additionally, success criteria were shared with guest lecturers. For instance, one of my guest lecturers specialises in ethics and law in Australian dental and healthcare. Detailed success criteria helped define content expectations and provided an opportunity for the lecturer to review and edit for clarity and posterity. It also guided lecturers on expectation and allowed room for detailed clarification prior to running workshops.

These actions were performed for all DENT3000 and DENT7221 course objectives.

Patient Case Studies: Streamline and Simplify Course Content: Use AI to Situate Public Health and Professional Practice in Clinical Settings

Context

One of the greatest challenges of this course is to communicate the relevance of the content to clinical practice. Many of the students in the 2024 cohort consistently questioned the importance of range of topics we are required to cover to meet accreditation standards set by the Australian Dental Council. This concern was echoed in the 2024 SECaT survey, where a student recommended that future iterations of the course place greater emphasis on this.

Objectives

To create a set of inter-related patient case studies that can be incorporated into a course handbook and embedded as teaching material into different assets. The case studies incorporate the learning objectives and success criteria for the course/s.

Outcome

The use of case-based scenarios that adapted learning objectives and success criteria. Here is an exemplar of what I created. The following section briefly introduces the patients to the students. It provides some context, introduces the family, briefly describes each patient and finishes with some reflective questions that are aligned with the learning objectives. Some of the AI elements used to edit and modify the ‘patients’ can be found in section ‘How I used AI to bring patients to life’.

Reflections

While student feedback on the redesigned course was generally positive, several themes emerged from the 2025 cohort of DENT3000. Students highlighted the value of case studies, noting that they clarified how the modules applied to different clinical contexts:

International students emphasised the opportunity to learn about the Australian healthcare and dental care system, with one noting they had “gained more insights into Aboriginal and Torres Strait Islander peoples as an international student.”

Students also acknowledged the broader professional importance of the course:

How I Used AI to Bring ‘Patients’ to Life

Context

In 2015, I led a university funded teaching innovation project that created ‘virtual patients’ to integrate diverse elements of the Dentistry curriculum into patient-based scenarios. We developed scripted cases that were recorded and edited with practicing dentists playing their roles alongside hired actors. These primer videos were integrated across multiple courses and continue to be used by our course coordinators today.

Drawing on this earlier experience, I scaffolded learning objectives into the newly developed (2025) patient-centred case studies using them as a tool to bridge course theory with clinical practice (McLean, 2016).

Objectives

1. Create photo-realistic images of patients.
2. Create a short video of the patient/s.
3. Enhance patient case study.
4. Give the patient ‘character’.
5. Give the patient a voice.
6. Use the patient to illustrate real-world scenarios.

Reflections

Creating text to image and video prompts generates engaging content when compared with static visuals (Pyae, 2024). However, AI-generated visual content introduces new conundrums to include:

• Hallucinations where AI ‘makes up’ the missing pieces rather than admitting it doesn’t know (Pyae, 2024).
• Copyright issues: We don’t know the origin or consent procedures of the images used to train the data (Qadri et al., 2023).
• Breaching the rights of vulnerable populations such as children and First Nations People such as Indigenous cultural intellectual property and perpetuating biases and disparities (Qadri et al., 2023).

On the other hand, AI generated images that can help maintain privacy as readers and students are unlikely to identify the images as ‘real people’. This allows course creators license to incorporate sensitive details that are often excluded in case studies to preserve patient privacy and confidentiality. For students, these AI-enabled patient scenarios more closely align with real-world situations. However, current, text to speech models produce voice-overs that can sound robotic and unnatural (Pyae, 2024). Consequently, while this was an interesting tool to play with, students and staff noted the ‘robotic’ quality of audio during focus group sessions as in the 2025 SECaTs.

All the case studies in my program were grounded in real-world scenarios. Some were drawn directly from my own clinical experiences, while others were informed by my subject knowledge, additional research and oversight from guest lecturers.

When I first began experimenting with AI-generated images and voices, I had not considered the ethical dimensions, regarding where the data came from and consent. Since the introduction of tools such as ChatGPT, I have become more aware that some companies have shifted their practices to incorporate ethical use by exclusively training models on data from consenting individuals who are willing to have their information shared. This is something I became aware of while writing a scoping review on the use of AI in dental and healthcare (Weerakoon et al., 2025). Consequently, the ethical use of AI will continue to be an ongoing area of exploration I plan to conduct with my students.

Create Content that Aligns with UDL Principles

Inclusion is also a subjective matter: what one student perceives as inclusion, might not be for another; there is unlikely to be one solution. (Tai et al., 2021; p.1936)

Context

Dentistry courses aim to equip graduates with knowledge and skills needed for practical hands-on outcomes. However, this knowledge is often delivered in micro-learning formats; small bite-sized segments that fragments overarching understanding and context. Such an approach risks creating knowledge silos which can take considerable time and clinical experience to integrate into a broader more holistic understanding of dentistry (Kodali et al., 2024).

Typically, clinically focused courses often emphasise memorisation of definitions, material properties, and medication names and dosages, sometimes at the expense of developing a broader, big-picture understanding. This narrow focus can de-skill students in reading and writing as they prioritise the acquisition of piece-meal knowledge (El-Kishawi et al., 2020). By contrast, reading and writing intensive courses such as DENT3000 and DENT7221 are often perceived as overwhelming. The reliance on reading material frequently drawn from government publications can appear verbose and demanding, and requires higher order cognitive skills to conceptualise, synthesise and generate meaningful discussion. For some students, a lack of prior skill development in those areas is challenging. It can increase the divide between those able to adapt to conventional academic methods (Ralabate, 2011) and those who face additional barriers such as students from CALD backgrounds or those experiencing neurodivergence (Tai et al., 2021). Student feedback can help identify differences in learning needs (Ralabate, 2011).

DENT3000 (2024) Pre-AI Feedback with Preference for Less Reading Material

DENT3000 (2024) Pre-AI Feedback with Preference for More Reading Material and Less Visuals

Objective

To create content that aligns with UDL principles (Ralabate, 2011) by converting required and recommended readings into podcasts and developing multimodal formats to improve student learning. While the dominant feature of this section is the creation and publication of a podcast series, alternative approaches were also considered to enhance accessibility and engagement.

Reflections

The podcasts were well received by many students, particularly for their accessibility and flexibility:

For many, podcasts were valued as time-efficient, flexible, and engaging. However, some students expressed a stronger preference for visual or interactive resources.

Others voiced dislike of podcasts altogether, emphasising the importance of offering alternatives:

Podcasts successfully supported many students by offering flexible, multimodal access to core content. However, student responses highlighted diverse preferences, some valued auditory learning for its efficiency, while others required visual or textual alternatives to remain engaged. This reinforces the importance of UDL-aligned design that provides multiple representations of content including transcripts, interactive modules, and visual media so all students can engage in ways that suit their learning needs.

Create Teaching Content

Use AI to structure content for lecture slides.

Reflection on Using AI to Structure Teaching Content

I used AI to help create the basic structure of a lecture on the Australian oral health system, a topic I consider to be conceptually challenging. This health system involves a variety of fragmented stakeholders and funding bodies to challenge the creation of an overarching ‘system’. In this instance, I found it useful to use AI to help me organise and categorise ideas. It reduced my cognitive load by providing me with a framework I could build on, rather than requiring me to start from scratch. However, AI-generated lecture slides are not without fault. Key issues I observed are:

• Images matched to AI-created content can be contextually inappropriate.
• AI includes adequate information, yet requires human engagement to edit for clarity, improve detail and enhance context.
• AI is not a content expert. It is also important for the teacher to be knowledgeable on the topic. For instance, in the section under ‘Commonwealth Government Involvement’, the slides generated referred to the ‘Teen Dental Plan’ that was replaced by the Child Dental Benefit Schedule.
• Some students perceived AI-generated lecture content to be ‘robotic’, and express dissatisfaction.

Use AI to Create Actionable Tutorial Documents.

The 2024 cohort expressed the need for targeted guidance to help guest lecturers align content with course learning objectives.

In response to this request, I created online tutorial guides for all guest lecturers. An exemplar is used to illustrate the process and outcome.

Reflections

I found the creation of tutorial documents helpful for both students and the guest speakers. Guest speakers often rely on re-purposed content from other presentations, which can lack the relevance or alignment with course needs. In the past, I had provided only the relevant course objective, often expressed as a single broad sentence. Developing a structured tutorial guide changed this dynamic. Contributors did not have to prepare content and could freely discuss topics following the guide. This also meant that students could follow the tutorial document closely. All the online tutorials were hosted on Zoom which has a feature that uses AI to transcribe the content which helps those who prefer to read rather than listen, to absorb content in their preferred mode.

Overall, the quality and relevance of tutorial content in 2025 improved significantly and many of the students responded positively to this change.

Assessment Support and Feedback Innovation

Creation of Assessment FAQs and Exemplars

In 2024, I responded to student requests by providing an assessment exemplar for DENT3000. Student use of the exemplar varied, some treated it as a benchmark to replicate, while others drew on it to extend their engagement with the rubric and produce advanced submissions. This reflects the literature that combining exemplar with rubrics can enhance student engagement and performance (To et al., 2022). Students who produced ‘outstanding’ work often used the exemplar as a springboard to ask exploratory questions from me to deepen their understanding of the assessment criteria.

A recurring challenge for me was the large volume of repetitive questions asked by students to clarify assessment requirements. Despite attempts to re-direct students to the discussion boards in our LMS, many students did not participate nor access the discussion boards in a timely manner. As one noted:

In response, I created Frequently Asked Questions (FAQ) documents for each assessment task and uploaded them into relevant assessment folders. These documents reduced email traffic and post workshop queries, and could be updated dynamically so all students accessed consistent information in one location.

Integration of an AI Declaration on Assessment Coversheets

Assignment coversheets are widely used to reinforce academic integrity (Gonsalves, 2025). Prior to 2025, however, they contained no reference to AI use. As my course profiles and task sheets had to be finalised before UQ released central guidance in March 2025, I developed my own AI declaration prompt and added it to the School of Dentistry assignment coversheet on 4 Feb 2025.

This encouraged some students to disclose AI use.

It is likely that some withheld disclosure for fear of retribution (Gonsalves, 2025). The link between disclosed AI use and assessment outcomes remains an area for further investigation.

Use of AI in Generating and Editing Feedback

Providing personalised feedback within my 0.4 role was a significant workload challenge. In 2025, across DENT3000 and DENT7221, I marked approximately 232,000 words and 5.5 hours of video assessment. To accelerate the feedback process, I experimented with ChatGPT in different ways. In 2024, I uploaded rubrics for a video-based task to generate boilerplate comments. In 2025, I extended this by drafting personalised feedback myself, then using ChatGPT to edit and refine for clarity. I also collated all feedback into a single thread to generate cohort-level insights.

This approach improved efficiency and allowed me to deliver detailed, personalised feedback at scale. However, student responses to AI -mediated feedback were mixed. While some appreciated the detail and the timeliness, others expressed discomfort.

Reflection

From a SoTL perspective, these practices illustrate the potential and tensions of integrated AI into assessment support. The documents created based on FAQs and exemplars addressed persistent issues of transparency and access, while the AI declaration foregrounded responsible use of emerging technologies in line with institutional policy. AI–supported feedback offered efficiency and scalability, but also raised questions about authenticity, trust and the role of humans in learning (Gonsalves, 2025).

Conclusion and Future Directions

Reflecting on the 2025 iterations of DENT3000 and DENT7221, did I achieve my overarching aim? I believe I did. Students were provided with multiple opportunities to contextualise public health and professionalism within clinical practice through varying modalities. I was also able to draw on AI as a co-creative teaching assistant to reduce cognitive load and save time. From a SoTL perspective, the integration of AI was not an end but a means to enhance teaching presence, diversify modes of engagement and support student learning. Student feedback demonstrates both the potential and he limitations of AI–enabled approaches reinforcing the need for careful design, transparency and alignment with pedagogical intent.

For future versions, I plan to build on these foundations by:

1. Expanding multimodal learning opportunities centred on podcasts and related readings in line with UDL principles.
2. Develop a comprehensive course handbook using the assets created throughout 2025 into a cohesive resource.
3. Evaluate student use of AI in assessment tasks and compare this with assessment outcomes to better understand the relationship between declared AI use, performance, and learning processes.
4. Embed teaching about ethical use of AI in practice and assessment (Weerakoon et al., 2025) with an emphasis on
5. Accuracy and trust in AI outputs, and
6. The limitations of AI applications in clinical contexts.

References

Abdul Kadir, N., & Schütze, H. (2022). Medical educators’ perspectives on the barriers and enablers of teaching public health in undergraduate medical schools: A systematic review. Global Health Action, 15(1), 2106052. https://doi.org/10.1080/16549716.2022.2106052

AHPRA (2024). Meeting your professional obligations when using artificial intelligence in healthcare. Australian Health Practitioner Regulation Agency. https://www.ahpra.gov.au/Resources/Artificial-Intelligence-in-healthcare.aspx

El-Kishawi, M., Khalaf, K., Al-Najjar, D., Seraj, Z., & Al Kawas, S. (2020). Rethinking assessment concepts in dental education. International Journal of Dentistry. 8672303. https://doi.org/10.1155/2020/8672303

Flodén, J. (2017). The impact of student feedback on teaching in higher education. Assessment & Evaluation in Higher Education, 42(7), 1054–1068. https://doi.org/10.1080/02602938.2016.1224997

Flaitz, C. M., Carlin, N., Shepherd, B. W., McWherter, J. A., Bebermeyer, R. D., Walji, M. F., & Spike, J. (2011). The journey beyond silos. Teaching and learning interprofessional ethics at UTHealth. Texas Dental Journal, 128(8), 716.

Gonsalves, C. (2025). Addressing student non-compliance in AI use declarations: implications for academic integrity and assessment in higher education. Assessment and Evaluation in Higher Education, 50(4), 592–606. https://doi.org/10.1080/02602938.2024.2415654

Hashem, R., Ali, N., El Zein, F., Fidalgo, P., & Abu Khurma, O. (2024). AI to the rescue: Exploring the potential of ChatGPT as a teacher ally for workload relief and burnout prevention. Research and Practice in Technology Enhanced Learning, 19, 23. https://doi.org/10.58459/rptel.2024.19023

Holtzman, J. M., Elliot, N., Biber, C. L., & Sanders, R. M. (2005). Computerized assessment of dental student writing skills. Journal of Dental Education, 69(2), 285–295. https://doi.org/10.1002/j.0022-0337.2005.69.2.tb03915.x

Keane, E., & Ní Labhraínn, I. (2005). Obtaining student feedback on teaching and course quality (Briefing Paper No. 2). Centre for Excellence in Teaching and Learning (CELT). National University of Ireland, Galway.

Kodali, M. V. R. M., Kodali, U. S., Gadicherla, S., Smriti, K., Singh, A., & Khurshid, Z. (2024). The role of soft skills in dental education: Challenges and importance. European Journal of Dentistry, 19(3), 851–859. https://doi.org/10.1055/s-0044-1791938

Malik, M. A. & Amjad, A. I (2025) AI vs AI: How effective are Turnitin, ZeroGPT, GPTZero, and Writer AI in detecting text generated by ChatGPT, Perplexity, and Gemini? Journal of Applied Learning & Teaching, 8(1). https://doi.org/10.37074/jalt.2025.8.1.9

McLean, S. F. (2016). Case-based learning and its application in medical and health-care fields: A review of worldwide literature. Journal of Medical Education and Curricular Development, 3, JMECD.S20377. https://doi.org/10.4137/JMECD.S20377

Nassar, A. K., Waheed, A., & Tuma, F. (2019). Academic Clinicians’ Workload Challenges and Burnout Analysis. Curēus (Palo Alto, CA), 11(11), e6108. https://doi.org/10.7759/cureus.6108

Pyae, A. (2024). Understanding student acceptance, trust, and attitudes toward AI-generated images for educational purposes. arXiv preprint arXiv:2411.15710. https://doi.org/10.48550/arxiv.2411.15710

Qadri, R., Shelby, R., Bennett, C. L., & Denton, E. (2023). AI’s Regimes of Representation: A Community-centered Study of Text-to-Image Models in South Asia. Proceedings of the 2023 ACM Conference on Fairness, Accountability, and Transparency, 506–517. https://doi.org/10.1145/3593013.3594016

Ralabate, P. K. (2011). Universal design for learning: Meeting the needs of all students. ASHA Leader, 16(10), 14–17. https://doi.org/10.1044/leader.FTR2.16102011.14

Tai, J., Ajjawi, R., & Umarova, A. (2021). How do students experience inclusive assessment? A critical review of contemporary literature. International Journal of Inclusive Education, 28(9), 1936–1953. https://doi.org/10.1080/13603116.2021.2011441

To, J., Panadero, E., & Carless, D. (2022). A systematic review of the educational uses and effects of exemplars. Assessment and Evaluation in Higher Education, 47(8), 1167–1182. https://doi.org/10.1080/02602938.2021.2011134

University of Queensland, Institute for Teaching and Learning Innovation. (2025). Course teaching feedback. https://itali.uq.edu.au/advancing-teaching/evaluation-teaching/course-teaching-feedback

Weerakoon, A. T., Girdis, T., & Peters, O. (2025). Artificial intelligence in Australian dental and general healthcare: A scoping review. Australian Dental Journal. https://doi.org/10.1111/adj.70000

Xiao, P., Chen, Y., & Bao, W. (2023). Waiting, banning, and embracing: An empirical analysis of adapting policies for generative AI in higher education. arXiv. https://doi.org/10.48550/arXiv.2305.18617