Here is a review of the AI in Higher Education Symposium at the University of Sydney on Friday, February 7, 2025. These are all the sessions that I attended (and there were lots!). I have incorporated the information from my notes and the conference program and used NotebookLM to assist in the compilation.
Introduction
The AI in Higher Education Symposium, held at the University of Sydney, brought together educators, researchers, and technologists from across Australia, New Zealand, and beyond to explore the transformative potential of Artificial Intelligence in tertiary education. The symposium fostered critical discussions on AI’s role in assessment, curriculum design, and pedagogy while addressing ethical considerations and practical implementation strategies. Attendees shared insights from pilot programs, research findings, and innovative teaching practices, all aimed at empowering educators and informing institutional approaches to AI integration. The event served as a valuable platform for collaboration, knowledge sharing, and envisioning the future of AI-enhanced higher education. This review covers 15 presentations from the symposium, drawing on key themes and insights presented.
1. A holistic approach to assessment for meaningful Al integration
Clara Sitbon and Kria Coleman (The University of Sydney). Sitbon and Coleman presented a compelling case for holistic assessment design when integrating AI. Using a custom AI agent called ClaraBOT; their approach transformed assessment in advanced French to mirror authentic language revitalisation projects. Students engaged with AI throughout their research journey, from task comprehension to creative problem-solving. Regular consultations ensured meaningful learning outcomes. The key takeaway was to consider what assessment elements need to be secure versus open to AI use, breaking assessments into parts with secure components like oral presentations.
2. Rethinking the curriculum of a whole degree in the time of AI
José-Miguel Bello y Villarino (The University of Sydney. Bello y Villarino discussed the comprehensive redesign of the University of Sydney Law School’s LLB and JD programs, focusing on integrating AI technologies into legal research training while maintaining academic rigour. This initiative concentrated on Course Learning Outcome 3, emphasising ‘research’ and aligning it with program-level learning outcomes. He highlighted the importance of general AI training for legal research, considering that most legal firms already use AI. The secure element of their assessment included a three-hour handwritten examination.
3. “Doctor ChatGPT, I Presume?” – A Collective Journey Through the AI Jungle
Henrik Dindas and Frank P. Schulte (FOM University of Applied Sciences, Germany) detailed an experiential ‘learning safari’ designed to immerse educators in hands-on AI exploration across key teaching domains, from lesson planning to assessment feedback. Participants developed practical competency with generative AI while critically evaluating its pedagogical applications. The presentation underscored the importance of applied approaches to AI training, especially considering the EU AI Act, which necessitates AI oversight skills.
4. Empowering Educators, Informing Institutions: Collaborative GenAI Trials in Higher Ed
Antony Tibbs (Edith Cowan University). Tibbs presented on collaborative GenAI trials in higher education, emphasising their value in supporting and leveraging key projects across institutions. He discussed a hack-a-thon at ECU and the formation of a community of practice using custom GPTs and learning agents for the Scholarship of Teaching and Learning (SoTL) and ethics research. Gathering valuable data from these trials, which put theory into practice, was a crucial outcome, allowing participants to discover the affordances of new technologies together.
5. Leveraging a multi-agent approach to enhance educator efficiency
Cory Dal Ponte (University of Melbourne). Dal Ponte explored enhancing educator efficiency using a multi-agent approach. They integrated AI into the ADDIE framework, using a backward design process to create custom GPTs. The presentation highlighted the creation of a learning design assistant with prompts for role, task, and format, emphasising granularity. Different bots were used for specific tasks, resulting in efficiency gains. This approach acknowledged that SMEs often lack an understanding of educational processes, thus requiring tailored AI assistance.
6. Generative AI as a tutor in the classroom to improve student understanding and provide teaching insights
Andrew Bartlett (The University of Sydney). Bartlett discussed using Cogniti and AI tutors to enhance student understanding. Cogniti was used to develop assessments, such as presentations, with an iterative process for agent development. This project highlighted the potential of AI tutors to provide teaching insights and improve student outcomes through personalised learning experiences.
7. Digital Sponsors: Transforming Project Management Assessments with AI Chatbots
Mark Freeman, Carmen Vallis, Enosh Yeboah (The University of Sydney). Freeman, Vallis and Yeboah shared insights into transforming project management assessments using AI chatbots with a cohort of 1000 students. The project focused on adapting to changes in the workforce, a crucial transferable skill. They employed scenario-based learning (SBL) to simulate real-world situations. The chatbot was designed to provide a drip feed of information, acting as a client for a project management project. Students submitted time-stamped logs of their conversations with the bot, promoting adaptation in project management.
8. Creating AISLA: An AI Student Learning Application for our course, “AI for Educators.”
Kathryn MacCallum, Kieran Williamson, David Rose, David Parsons (University of Canterbury). MacCallum, Williamson, Rose and Parsons presented on AISLA, an AI Student Learning Application for an “AI for Educators” course, designed as an asynchronous short course for AI literacy. The application used personas, such as sceptical colleagues, to engage educators. Chatbots were integrated into the LMS, but the project extended beyond chatbots, considering that not all learning objectives involve conversation. It was highlighted as a pilot study to explore the possibilities of AI in educator training.
9. Building AI Literacy: An Adaptive Self-Assessment Tool for Student Readiness
Samantha Clarke, Lucas Wright, Simon Bates, Eszter Kalman, Jessica Frawley, Leah Wafler, Donna Langille, Heather Berringer (The University of Sydney and the University of British Columbia). Clarke et al. discussed building AI literacy through an adaptive self-assessment tool for student readiness, addressing common challenges across different contexts. The project used co-design principles, partnering with students to develop the chatbot in 80 days. Aligned with international competencies like UNESCO, the tool was continually designed through testing, especially with student input. The AI chatbot prompts students to use GenAI, featuring 14 questions to test AI literacy adaptively.
10. Developing AI-Guided Learning Modules: Insights from the GAIL Experience
James Bedford and Patrick Tran (UNSW College and UNSW). Bedford and Tran presented on developing AI-guided learning modules, drawing insights from the GAIL (Generative Artificial Intelligence Literacy) experience. They used GenAI to teach GenAI, creating a custom experience with adaptive learning pathways. The method was Socratic, gamified within guardrails, and available in the custom GPT store.
11. Enhancing Accessibility, UX and build efficiency in Learning Design through Custom GPTs
Jesse Keenan, Callum Fitzpatrick, and Fabian Dal Forno (Online Education Services). Keenan, Fitzpatrick, and Dal Forno from OES discussed enhancing accessibility and UX and building efficiency in learning design through custom GPTs. They highlighted the use of Enterprise Open AI, with custom GPTs visible across all teams and an AI working group ensuring consistency. One custom GPT created quizzes, converting them into LaTEX and HTML, while another generated text alternatives to images for accessibility, improving efficiency for the 72 learning designers.
12. From Accounting to Business, Communication, Engineering, Health, and beyond: The Impact of GenAI in UniSA Online’s Curriculum
Rebecca Godwin (UniSA Online). Godwin explored the impact of GenAI across various disciplines at UniSA Online, emphasising an industry focus. A task discussed involved using ChatGPT to write feature stories for a news outlet. This presentation showcased the broad applicability of GenAI in curriculum design and assessment.
13. Beyond ChatGPT: Understanding Chinese AI Tools and Their Implications for International Students in ANZ Higher Education
Pedram Nourani & Anthony Ryan (University of Waikato). Nourani and Ryan addressed the importance of understanding Chinese AI tools and their implications for international students in ANZ higher education. They suggested designing assessments so students can use any AI tool with critical documentation. The focus should be testing the learning objective, not the tool itself. They discussed the differences between Chinese AI tools versus Western tools, advocating for AI agnosticism and involving international students in the discussion.
14. AI-Powered Pipeline for Personalised Exam Preparation and Targeted Learning Intervention in Large Cohorts
Angela Sun and Helen McGuire (The University of Sydney). Sun and McGuire presented an AI-powered pipeline for Personalised Exam Preparation and Targeted Learning Intervention in Large Cohorts. They used scenario-based questions in exams, with a Cogniti-based bot for preparation. The data extracted from student responses was real, from past exam questions. The project measured using Cogniti with exam outcomes, assessing the tool’s impact on a closed-book, scenario-based assessment task.
15. EweSYD: Immersive GenAI digital simulations for teaching complex physical processes.
Mike Seymour (The University of Sydney). Seymour introduced EweSYD, an immersive GenAI digital simulation for teaching complex physical processes. This project, originating from the MOTUS lab at USyd, used AI and ML, not just LLMs, to determine if a sheep is pregnant using Apple Vision. It highlighted innovative uses of technology beyond traditional language models in education.
Conclusion
The Symposium highlighted AI’s diverse and rapidly evolving landscape in tertiary education. Key themes emerged across the presentations, including the importance of holistic assessment design, the need for AI literacy among educators and students, and AI’s potential to enhance learning outcomes through personalised and adaptive approaches. The integration of AI into curriculum design, as demonstrated by the University of Sydney Law School’s comprehensive redesign, signals a significant shift towards preparing students for an AI-driven future. Furthermore, the ethical considerations and the need for critical evaluation of AI tools were consistently emphasised, reflecting a commitment to responsible AI implementation. Ultimately, the symposium underscored the transformative potential of AI in higher education while highlighting the importance of thoughtful, ethical, and pedagogically sound integration strategies. The collaborative spirit and wealth of shared insights will undoubtedly contribute to the continued advancement of AI in education across the region.
Leave a Reply