Written exams, presentations and essays have been a reliable and scalable cornerstone of how universities assess who knows what. But artificial intelligence tools like ChatGPT and Claude are disrupting that model, forcing business schools to revisit how they preserve academic integrity.
Research from Ivey Business School’s Tiffany Bayley, Assistant Professor of Management Science, Kyle Maclean, Assistant Professor of Management Science, and Tessa Weidner, Director of the EdTech Lab, conducted during the pandemic and equally applicable in the era of AI suggests business education may need to look at a more analog form: the oral exam.
While teaching online during the early days of COVID-19, Bayley, Maclean, and Weidner explored the use of Concurrent Video-Based Oral Exams — or ConVOEs — an assessment approach where students simultaneously record and submit short video responses to exam questions, rather than writing. They published their findings in the paper, Back to the Future: Implementing Large-Scale Oral Exams.
“We understood that the assessment had to be at the same time for every single student in the class to avoid the temptation to share questions,” says Bayley. “We wanted to actually hear in their own words what they were understanding from the class.”
Six years later, the shape of the problem has shifted, but the research has become even more relevant.
“Back then, it was by asking other people and Googling,” says Maclean. “Now it’s putting questions in ChatGPT.”
According to a report from KPMG, in 2025, 73% of Canadian students used generative AI tools to complete their schoolwork, a sharp rise from 59% in 2024. While the data shows students are embracing new technologies, two-thirds of those surveyed say they are not learning or retaining as much knowledge as a result of their AI usage.
The researchers argue that oral assessment helps preserve academic integrity, while also giving the students a chance to work on another vital skill: communicating and defending their ideas in real-time.
The skills gap hiding in plain sight
Studies of thousands of job postings and graduate program curricula found that communication skills rank among the most sought-after competencies employers want, yet are underemphasized in business school.
In most business analytics courses, communication is an afterthought — a modest slice of the grade on a group project, secondary to the technical work that actually gets taught. According to the paper, two-way communication, listening, responding, conversing under pressure, is a key differentiator for candidates in professional settings.
“It’s no different than a client presentation,” says Maclean. “If a client comes back to you and says, Oh, you said that’s a negative binomial, what does that mean? And you say ‘I don’t know.’ Your whole presentation has gone downhill. You’ve lost your credibility.”
Everything old is new again
The ConVOE model was designed to solve this gap. By having every student submit simultaneously through their learning management platform, the team built an assessment that was individual in nature but scalable in practice.
“Students did find it stressful,” says Bayley. “But they could also see the benefit. It really made them think about how to speak normally about the stuff that they’ve learned.”
The faculty also found the experiment insightful. Listening to students work through the same concepts in their own words exposed moments where the instruction was unclear. “It helped me surface things I thought I could do better,” says Bayley.
It was valuable data that no written exam could capture, a diagnostic tool as useful for instructors as it was for students.
That insight points to something broader. Oral assessment isn’t just an academic integrity measure. Used well, it’s a feedback mechanism, a communication development tool, and a closer proxy for the environments graduates will actually work in.
The question isn’t whether it’s worth doing. It’s how to make it feasible.
Incorporating oral assessments effectively
For instructors willing to experiment, the researchers are careful to say the point isn’t to replicate what they built in 2020. That model was built for a specific moment, and that moment has passed. What remains useful isn’t the recipe so much as the reasoning.
1. Start with what you already have.
The instinct when facing a new pedagogical challenge is to find a new platform or format. Weidner recommends resisting it. The infrastructure for oral assessment, whether that’s a classroom, a learning management system, or a scheduled Q&A, already exists in most teaching environments. For the ConVOE model, the researchers used their school’s pre-existing learning management system.
2. Think in parallel, not in sequence. The reason oral exams fell out of favour at scale is simple arithmetic: evaluating students one at a time is prohibitively time-consuming. The solution isn’t to abandon the format — it’s to design around the constraint. That might mean simultaneous video submissions graded asynchronously, or splitting a large class into smaller groups assessed concurrently by teaching assistants. The goal is the same: individual accountability, without the bottleneck.
3. Find the moment that demands a real answer.
The format matters less than the question it creates: can this student defend their own thinking? That moment can be engineered in a lot of ways — a structured Q&A after group presentations, a brief oral component at the end of a project, a single follow-up question on a written exam. Maclean’s recent restructure of group presentations, five minutes to present, twenty minutes of probing questions, required no new technology or infrastructure, just a willingness to make the conversation the assessment.
“The oral exam isn’t meant to be this catch-all,” says Bayley. Ultimately, it’s about instructors thinking about what they’re actually trying to assess.
Want to learn more? Read Bayley, Maclean and Weidner's full paper, Back to the Future: Implementing Large-Scale Oral Exams, published in Management Teaching Review.
