Not long after our School’s Quality Assurance meeting (i.e. the meeting where we finalise our students’ grades), a group of colleagues and I headed to a nearby pub for what could best be described as a ”debrief beer’. Unsurprisingly, as is characteristic when you gather a group of academics who care about education, the conversation eventually turned to assessment. One comment from a colleague particularly stuck with me. They observed that there seems to be an increasing temptation for students to shortcut learning. This really got me thinking.
I am not an expert in broader sociocultural trends, nor am I attempting to provide an authoritative commentary on the changing nature of education. However, especially across higher education, concern is growing that students can use AI to complete assessments without engaging deeply with intended learning. Discussions often focus on academic integrity and assessment in the era of AI, but the practical question is how to provide credible evidence of learning that suits the competency being assessed.
From artefacts to evidence
Reports, exams, essays, presentations and design submissions are established assessment methods. A report is assumed to represent understanding, a design submission to represent engineering judgement. and presentation to represent communication capability.
Historically, as student numbers and programs expanded, universities needed approaches that could be administered consistently, moderated fairly and managed within finite resources. These methods enabled assessment at scale, with reasonable confidence in achievement. They also rely on artefacts as proxies for learning: a report for understanding, a design submission for engineering judgement, a presentation for communication capability. However, traditional examinations have well-documented limitations in assessing contextualised and authentic application of knowledge.
These artefacts remain valuable, but their suitability depends on the competency being assessed. Where an outcome includes applying knowledge to an unfamiliar problem, explaining assumptions or responding to evidence, a finished product may not, on its own, show all the thinking involved. The question is not whether the artefact has become worthless, but whether it alone provides the confidence we need. Assessment should be matched to the competency, rather than expecting one format to demonstrate every outcome. No single format suits every outcome or cohort; a method may provide strong evidence of one capability and weaker evidence of another.
Oral defences, live demonstrations and viva-style assessments are often presented as innovative responses to contemporary challenges, yet they are among the oldest university assessments. Before universities managed hundreds or thousands of students across programs, understanding was demonstrated through explanation, defence and discussion. We could infer that oral assessment became less common because it seemed difficult to scale, not because it lacked educational value.
Making student thinking visible
In my own teaching within biomedical engineering, I have used oral assessment as a secure component of several take-home assessments.
- In BMET2400 Biomechanical Design, students create engineering designs and technical drawings using CAD software before discussing and defending their design decisions during an oral assessment.
- In BMET3961 Biomaterials, students investigate the materials used in medical devices and subsequently answer questions about them without notes.
- In BMET3971 Tissue Engineering, students read seminal papers and then participate in oral discussions designed to probe their understanding and reflections.
Across these examples, I have found oral assessment valuable not because it identifies misconduct or verifies authorship, but because it makes student thinking visible. Students are asked to explain assumptions, justify decisions, interpret findings, and respond to probing questions. The resulting conversations often reveal dimensions of understanding that are difficult to infer from a written submission alone. More importantly, they provide confidence that students genuinely own the knowledge and ideas represented within their submitted work.
These assessments use discussion to gather evidence that complements the submitted work: students explain design decisions, respond to questions about materials, or discuss and reflect on research. The format is useful because it lets students demonstrate capabilities the unit is trying to develop. This does not mean every outcome requires a viva; it means being clear about what students should demonstrate, then choosing evidence that can show it.
Designing for delivery
Consider a cohort of 300 students. Even a brief ten-minute conversation with the cohort represents approximately 50 hours of assessment time before accounting for preparation, moderation, feedback, scheduling, administration, and benchmarking between assessors. Having twice conducted individual oral assessments for close to100 students during Semester 1 2026, I can attest to both their value and the demands they place on me.
Distributing the assessment across multiple tutors and benchmarking can help, but it takes careful preparation. Coordinators may spend dozens of hours designing, refining, calibrating, and reflecting on an assessment. Tutors understandably engage with the assessment at varying depths and for varying lengths of time. As the number of assessors increases, achieving consistency and confidence becomes more difficult, even with substantial benchmarking and moderation. Individualised assessment also requires suitable physical spaces and careful scheduling.
This implementation work is not unique to oral assessment. Any new assessment takes time to clarify outcomes, design and refine tasks, develop marking guidance, prepare staff, enter grades and provide feedback. Calibration and review continue as it is used. Consider these demands alongside benefits: oral conversations can provide richer evidence, while written tasks may suit other competencies. The choice should follow learning outcomes and available resources.
So what should we do, collectively?
The temptation to shortcut learning will remain, and generative AI has made some shortcuts more visible. The challenge is deciding what evidence gives us confidence that learning has occurred. Oral assessment is not the answer, but it highlights the tension between gathering strong evidence and assessing at scale. For individual educators, this means asking where understanding becomes visible within a unit. Are students primarily rewarded for producing artefacts, or can they also explain, defend, apply and justify their thinking? What evidence best matches each competency? The answer may be an oral discussion, written explanation, demonstration, staged task or combination. The goal is to make the relationship between learning outcomes and assessment evidence explicit.
For many years, universities and academics have balanced assessment approaches that scale efficiently with reasonable confidence in the learning they certify. Shortcuts invite us to examine that balance. Assessment artefacts became dominant partly because they are practical at scale. As educators, we should ask what evidence would convince us that students have achieved intended outcomes. My experience with oral assessment has shown me the value of richer evidence and the work required to obtain it; our choices should reflect both the learning we want to see and the conditions needed to assess it well.