Introduction
In July of ‘26, I attended the Higher Education Research and Development Society of Australasia (HERDSA) conference held at the Stephen Riady Centre in the National University of Singapore. The main goal of the program is to promote the development of higher education policy, practice and the study of teaching and learning in tertiary and higher education. In this mini essay, I will go over some thoughts I’ve had attending my first conference, and some takeaways I’ve had from listening to the speakers.
The conference is structured to have concurrent sessions with multiple presentations in different venues at the same time. Attendees choose which session to attend during each time slot. I attended talks that interested me, particularly those on how educators use AI as a teaching tool and how they managed students relying on it for thinking or assessments. Overall, I listened in on ten talks and derived some kind of takeaway from nine of them.
Evaluating live online quizzing as a response to AI misuse in tertiary assessment* by Dr Lee Harrison, University of Waikato
In this talk, the speaker evaluated the use of a Moodle-timed assessment with forward-only navigation during his tutorial classes. Each assessment would have approximately ten questions based on the previous week’s lecture content. Students were given 60 seconds per question to key in their answers (MCQ and short-answer responses) on their devices for online submission. Questions were displayed on the classroom projector, ensuring fairness. All students were allowed to refer to one A5, double-sided, hand-written cheat sheet during quizzes. The weightage of the quizzes were low (2-3% per quiz) to reduce anxiety and nervousness during the assessment. To make up for this, in-person assessments, like presentations and viva voces, were weighted higher to make up for the low overall weightage of quizzes. There were no final examination components in the course.
The majority of student responses were positive and many preferred this assessment style over take-home assessments as the fixed time limit reduced procrastination and competing commitments. They also found the cheatsheet valuable, not mainly for use during quizzes, but because preparing it encouraged them to review the previous week’s content and synthesize the information by writing in on a piece of paper—an activity they would have not done for take home assessments.
Takeaway(s): In the past, my perspective of assessments was primarily viewing them as assessments for learning or assessments of learning; I had not thought of how assessments could be used as a means to encourage reviewing lecture content through the use of cheatsheets.
Skills taxonomies in higher education* by Ms Himani Chugh, The University of New South Wales
In this talk, the speaker highlighted the disconnect between the soft skills educators sought to impart to their students and the soft skills in-demand by the workforce. More notably, she noticed a lack of common language and terminology between the educators and employers which left students confused: many did not know what soft skills they’ve been taught and how they match up to the responsibilities required in job descriptions.
To combat this, the speaker listed nine specific “enduring human skills”:
- Collaboration
- Communication
- Creativity
- Critical Thinking
- Digital Literacy
- Organization
- Problem-solving
- Self-regulation
- Technical
Takeaway(s): Educators should explicitly articulate the intended learning outcomes (ILOs) of their course so that students are cognizant of which human skill is being developed. Moreover, educators should point out these skills in students and encourage open discussion to help students identify, articulate, and confidently communicate their own capabilities to future employers.
Thinking to learn: Designing assessments that preserve student owned thinking in the age of AI* ****by Dr Anne-Marie Chase & Dr Kelly Galvin, Swinburne University of Technology
In this talk, the speaker discussed the kinds of ideas and intentions educators should be aware of when planning their assessments. Rather than banning AI use altogether, they focused on how educators can create assignments that complement the use of AI while meeting ILOs. Additionally, where type and extent of AI use is concerned, they emphasized the need for a structured framework for students and educators to refer to when completing and designing assessments respectively—what they called “Constructive Alignment”.
Takeaway(s): We can’t put the GPT genie back in its bottle; setting clear boundaries on AI use in courses and designing assessments that complement AI use seem to be the more effective strategy. It might prove worthwhile to discuss on a AI usage scale for intended/allowed AI use in assessments with colleagues.
Evaluating an online viva voce capstone assessment to strengthen authentic assessment in the GenAI era* ****by Dr Lynda Hughes, Griffith University
In this talk, the speakers shared about how their university conducted a viva voce at scale: 803 students were assessed by 18 staff over an online viva voce. Recordings were taken for repeat viewing/later review by a second assessor to reduce bias. The main concerns brought up during the QnA were how to prevent the use of AI tools during assessments which the speakers admitted could not be prevented entirely.
Takeaway(s): Assessor calibration was highlighted as the most crucial factor to both upholding fairness and reducing inefficiencies caused by skewed results that demanded a second (or third) review. Clear rubrics and assessor training before the conduct proved to be a worthwhile investment as it saved time and reduced confusion during the conduct. Where AI use is concerned, see above takeaways on designing assessments that complement AI use rather than prevent them.
Practical pedagogy: The impact of human-AI feedback on collaborative dynamics in group writing tasks in higher education* by Dr Janice Wong, Singapore Institute of Technology
In this talk, the speaker shared data she collected during her teaching on how students viewed the usefulness and value of peer feedback, AI-generated feedback, and teacher feedback. Survey data was entirely self-reported and students opted-in to the study. During the course, students would receive feedback on writing assessments from their peers or from an LLM, but they always got teacher feedback so as to not compromise on teaching quality.
In her data, she found no significant difference in how students rated peer feedback and AI-generated feedback. This suggests that receiving peer feedback is on par with receiving AI-generated feedback, despite the former being established as an engaging learning process that encourages collaborative dialogue and develops critical thinking skills in students. It may also suggest that LLMs can serve as collaborative partners in the learning process when used appropriately.
Takeaway(s): Once again, I get the impression that designing assessments that complement AI use is more beneficial for both students and educators than trying to avoid AI use altogether. A general consensus I’m seeing is the importance of teaching students how to use LLMs, and then encouraging them to use AI appropriately to accelerate their learning.
Co-teaching, interdisciplinarity and learning-oriented assessment in higher education* by Dr Alfonso Lopez- Hernandez, Comillas Pontifical University
In this talk, the speaker discusses the benefits of co-teaching courses to educators, especially courses that are inter-disciplinary in nature. Co-teaching is defined designing and/or teaching a (part of a) course either alongside another educator or working under another educator to teach a course. The educators can be from the same or different disciplines.
Based on interviews and feedback from educators in Comillas Pontifical University, he found that educators who engaged in co-teaching focused more on task design and overall alignment in ILOs and assessment outcomes. In contrast, educators who did not co-teach tend to focus on their own discipline of study and paid more attention to evaluative judgment during assessments and consistency of feedback.
Takeaway(s): The main benefits of co-teaching seem to be that it generates discussion on (1) feedback practices and (2) assessment design: co-teaching keeps educators mindful of how their feedback and assessments should adhere to their ILOs. I suspect that a sign of poorly-designed co-taught courses is that the course content appears ‘messy’ or ‘irrelevant’ as students would not see the relevance between the ILOs and the assessments they complete or content they cover in class. It is important to be mindful of the ILOs of assessments so that they align with the course’s overall ILOs.
Assuring Learning in an AI-enabled era: An Institutional Assessment Framework* by Prof Maree Dinan-Thompson & Assoc Prof Nicole Masters, University of The Sunshine Coast
In this talk, the speaker discusses the issue of how to weigh assessment components in a course, especially when AI use is rampant and ubiquitous. There is much overlap between this talk and many of the prior ones on AI-use and assessment design. Some highlights are how the frequency of assessments for learning (AFL) and assessments of learning (AOL) need to be increased while keeping both at a low-weightage to reduce the impact of AI-use and preserve learning. Alternatively, one may reduce the weightage of AFL—while increasing their frequency—and increase the weightage of one or two in-person AOLs to encourage AI use in a low-stakes environment, while preserving the evaluative usefulness of tests and examinations.
Regardless of how educators choose to distribute the weights of their assessments, the most important factor is that they each assessment clearly maps to one or two ILOs, and that each ILO has both AFL and AOL components. Thus, regardless of whether students choose to use AI in their AFL components, they will still have enough opportunities to learn relevant concepts and skills without giving AI-users substantial edge over non-users.
Takeaway(s): Instructors need to be clear about which assessment maps to which ILO, especially during assessment design. This ensures that students have sufficient opportunities to learn the relevant concepts regardless of whether they use AI, and without giving those who use AI a substantial edge over non-users.
The Creativity Paradox: Why Our Feedback Holds Students Back* by Dr Abdul Razeed & Dr John Parker, University of Sydney
In this talk, the speaker warns educators about how the type of feedback they provide can drastically alter the thinking process adopted by students. They categorize feedback into three different types:
-
Direct feedback (DF)
Focuses on errors and inaccuracies; tells students how the work missed the rubrics requirements; the feedback generally comes from a place of lack and ascribes blame to the work
-
Facilitative feedback (FF)
Focuses on giving students opportunities to develop their own solutions, reflect on their ideas and justify their ideas; hands judgment back to them; uses prompts and scaffolding rather giving direct answers
-
Generative feedback (GF)
Focuses on points that were unique or underdeveloped to generate new questions, possibilities or future actions; the feedback generally comes from a place of growth and creates future learning rather than commenting on past performance
DF tends to discourage risk-taking and stifles creativity. It focuses students on a single task and removes student agency in they’re learning as they soullessly structure their submissions around meeting rubric requirements. On the other hand, FF and GF gets students to engage with the material in front of them with the goal of identifying areas to expand on rather than on faults with the work—resources to build from instead of looking for flaws to fix.
Takeaway(s): Try decoupling feedback from grades and rubrics, especially in low-stakes assignments (AFL). Adopt a growth mindset when writing feedback and focus on getting students to discuss and elaborate on any original ideas, even if it strays away from the rubrics. Where possible, increase the frequency of AFL while decreasing the weightage of individual AFL components so that students are less pressured about scoring highly.
”Interesting even at 9am”: how real-world storytelling and analogical scaffolding dismantle the “fluff” perception in communication-in-the- disciplines* by Mr Tony Han-Wah Goh, National University of Singapore
In this talk, the speaker reviews the course feedback from students on his engineering course over the past 4 years. The main reasons for the poor reviews during the course’s inception were primarily a disconnect between the ILOs and real-world application of the ILOs, leaving students frustrated over how the content taught to them was not applicable to their career or daily life. He highlights how the use of analogies, real-world examples and increased student participation drastically improved student receptiveness of the course content and teaching.
These methods fostered a friendly and engaging environment which stimulated thinking and participation in spite of the course being conducted at 9 am in the morning. Students could recognize, practice and independently apply the skills and frameworks taught to them on everyday examples which made them more receptive to its usefulness and aware of the teacher’s effectiveness in helping them connect these ideas. Only when the students both accepted and could apply these skills themselves did the educators go into actual engineering examples. This both accelerated the reviewing of more ‘dry’ content and reinforced students learning (AFL).
Takeaway(s): Use more analogies and real-world examples, especially in smaller tutorial groups. Increase student participation to foster a sense of ownership in they’re learning. Explicitly connect the course’s ILOs to everyday life, and reinforce learning through practice with familiar, real-world examples. Once students have built a strong foundation, transitioning to more abstract or “dry” examples becomes easier and more effective.