Reimagining Higher Education in the Age of AI
- Dr Sp Mishra
- Jul 18
- 5 min read
Updated: Aug 5
Episode #120 of the India Career Centre Podcast, with Professor Steve Joordens, University of Toronto
For decades, the value of education rested on a simple premise: knowledge was scarce, and institutions controlled access to it. That premise has quietly collapsed. Information is now available almost instantly, and largely at zero cost. So what happens to teaching, learning, and assessment once access to knowledge is no longer the bottleneck?
In Episode 120, I sat down for the fourth time with Professor Steve Joordens an educator, researcher, innovator in AI-enabled learning, and CEO of Talk Lab to work through this question in detail. Our conversation ranged from neuroscience to classroom design to the future of testing, and it left me more convinced than ever that we are at a genuine inflection point in how the world learns.
From access to verification in Higher Education
The starting point of our conversation was simple but far-reaching: if information is now a query away, what actually becomes scarce? Professor Joordens framed the moment in Kuhnian terms:
"I find this a very exciting time for education... I very much think of it as Thomas Kuhn thought about paradigm shifts in science, when we used to think the earth was the center of the world, and then at some point we had to kind of give up that way of thinking."
And on why simply outsourcing work to AI misses the point entirely:
"Simply asking AI to do stuff is not going to be a value. There's going to have to be humans in there working with the AI to produce good quality work and that's what we have to teach them how to do now."
Why multiple-choice testing may be on its way out
One of the most compelling threads in the episode was Professor Joordens' case against multiple-choice testing a format he has distrusted since his own days as a student. He put it bluntly:
"I think getting rid of multiple choice testing would be a huge benefit to education. So I'm sort of jokingly on what I call my death to multiple choice tour right now."
His argument is that multiple-choice was never really about measuring learning well; it was about measuring learning at scale, while oral assessment an expert probing what a student actually understands was always the better method that never scaled:
"The advantages of multiple choice, being able to assess learning at scale, but without the trade-off... maybe we can now say we can go back to the better way, and using AI, we can now do that at scale."
He shared data from his own classroom on how students actually respond to being orally assessed by an AI:
"There are a few there are a few like stressful or scary, but the real big ones are like fascinating, intriguing, different, unique... it's not like checking boxes on a multiple choice. It's like you want to tell it what you know."
For Professor Joordens, fixing assessment is "the first domino" in rebuilding education because how we assess determines what we end up teaching.
Teachers as coaches, not information dispensers
If knowledge is no longer scarce, the role of the teacher has to move with it. Professor Joordens described what he calls being a "workforce multiplier" an AI version of himself available to 1,800 students at once:
"It could do what I also call high touch at scale. So it could do one-on-one intense conversations, but literally it could do twenty of those at a time... and it works tirelessly."
He connected this directly to a broader shift the sector needs to make from evaluating the product a student hands in, to understanding the process they went through to produce it:
"We used to say hand in a report and we would just look at the report and it was all about the product. Now we're going to say create a report, but tell me about how you created that report... We care about the learning experience you went through to produce that thing and the skills that you've learned."
For a country like India, where student-to-teacher ratios are often heavily skewed, this model of "high touch at scale" is particularly relevant, AI-enabled personalisation that no single teacher could realistically deliver to hundreds of students individually. As Professor Joordens put it on the personalisation front:
"The AIs can do that. They can do high touch at scale. And so that makes them especially powerful to help students, even students with autism that might have special needs... It can really tailor the learning to the student in a way that teachers just cannot do."
The cost of convenience: what we may be losing
The conversation wasn't a one-sided celebration of AI. We also discussed a genuine trade-off: the deep, undistracted attention that comes from reading a physical book versus the fragmented attention that comes with learning through a phone or a screen. Professor Joordens was candid about this tension, even quoting his own sister:
"My sister sometimes says I spend my time using technology to help fix the damages caused by technology, which I think sometimes is true."
He described how he built an AI trained on his course readings to compete with his students' phones for their attention:
"I created an AI that knew those readings inside and out. I said, treat this like somebody sitting at your coffee shop that's also read the paper, that knows it well, and you can go and just talk to them about the paper... That was better competition for their phones."
But he was honest about what might be lost in the process:
"I kind of feel like the way we did it was the better way — where we would learn to shut out distractions and really focus our mind on something. That is a rare thing for students these days."
The real bottleneck: scaling good ideas
Perhaps the most grounded part of the discussion was about why these promising models haven't spread faster. It isn't a shortage of good ideas — Professor Joordens and his colleagues have already shown these approaches work at a small scale. As he described it:
"We need the research, we need the advocates, but you also need the businesses... If you think of multiple choice testing now, how many pieces are underneath that? There are businesses that support the multiple choice testing. There are all of these things in place. So if we want to overturn that, we have to be able to work at that scale. And that is going to be one heck of a ride."
The bottleneck, in other words, is ecosystem-level: research funding, institutional buy-in, regulatory bodies, and businesses like his own venture. Talk Lab is capable of delivering these tools reliably at scale.
Why this conversation matters
This episode isn't a discussion about whether AI belongs in education that debate is over. It's a serious, practitioner-level conversation about how to rebuild teaching, learning, and assessment once the old constraints no longer apply. For educators, policymakers, and parents navigating this shift in India and elsewhere, Professor
Joordens offers both a clear-eyed account of the challenges and a genuinely optimistic vision of what education could become.
Watch or listen to the full conversation Episode 120:
"Reimagining Higher Education in the Age of AI" — on the India Career Centre podcast.
Dr SP Mishra hosts the India Career Centre podcast, featuring in-depth conversations with educators, researchers, and industry leaders on the future of learning and work.
About the Author:
Dr SP Mishra is the Founder of India Career Centre, a Hyderabad-based career guidance, study-abroad consulting and educational research organisation. He writes and speaks on career decision-making, higher education, skills, employability and the future of work in India. Also hosts the India Career Centre Podcast. Connect on LinkedIn.
Comments