Tech won’t fix universities until they fix themselves

Prof. Soumitra Dutta
Former Dean, Saïd Business School, Oxford University
and SC Johnson School of Business, Cornell University

Technology can dramatically expand access to education. But its transformative potential will remain unrealized unless universities rethink their underlying model, argues Prof. Soumitra Dutta.

Gavin Allen: How should universities leverage new technologies?

Prof. Soumitra Dutta: Before thinking about technology per se, universities must first rethink their basic educational model. Each year, top US Ivy League schools admit a few thousand students. But demand for high-quality education far outstrips that supply. Technology allows universities to increase student access by a factor of 10, 100, or even 1,000 – with the same resource base.

But the culture in these universities is not designed for that kind of change. So, they're putting all this money into technology, but they don't see proportional productivity gains.

Gavin Allen: If I go from teaching 20 students to teaching 60, I may be more productive. But doesn’t education quality suffer?

Prof. Soumitra Dutta: That’s a myth. How much do you think the professor actually knows about each student’s learning and expectations in a classroom of 60 to 100 students? I can tell you, having been a professor myself, it’s very little. The amount of customization and adaptation professors provide is actually quite low.

There are exceptions, of course. Universities such as Oxford and Cambridge use a small group tutoring system that increases learning quality; that's one reason they’re ranked so highly. It’s inefficient, but it creates a lot of learning value.

Gavin Allen: At a very high cost.

Prof. Soumitra Dutta: So high, in fact, that many universities have started hiring PhDs to do the tutoring, even though this degrades teaching quality. Tutoring is so time-consuming that many professors pay to get out of it. So, yes: the Oxbridge system is not the most efficient. If we want to make it more productive, the whole system will have to change.

Gavin Allen: How can universities adapt to AI before it steamrollers them?

Prof. Soumitra Dutta: In a university, the faculty holds the power. In general, they have comfortable lives and every incentive to maintain the status quo. They will argue that if they have just 20 students per class, they can teach better. But what is their responsibility to society? To the world? Just to educate a few hundred students? Why not a few thousand?

These institutions should rewire their DNA so they can change the lives of more young people, too many of whom are getting left behind. We have to challenge fundamental assumptions, starting with the role of the professor. Universities have a teaching and a research mission. Top researchers in any university are a small fraction of the faculty pool. The rest churn out average papers, earn comfortable packages, and live happy lives. But do we need that? For many, the name of the game is to teach as little as possible and maximize research. The system must change.

Gavin Allen: Change how?

Prof. Soumitra Dutta: For the teaching mission, the answer is AI.

Today, you can build hyper-realistic avatars and personalized learning journeys for large numbers of students. But most universities will never do it because the faculty in power risk losing their jobs and their comfortable lives.

We need people who show empathy and act as counselors who build character, enable social development, and stimulate creative thinking. Instead, the ethos of universities has been diverted by an elite research metric.

Gavin Allen: So, emotionally intelligent facilitators to a far wider pool of students, with a greater emphasis on self-driven intellectual exploration?

Prof. Soumitra Dutta: Exactly. One of the most important goals of the university experience is social – a metamorphosis of the individual in a collective setting, as a teenager becomes a young adult during the undergraduate years. That’s largely been left to students to manage, with very little university investment.

Second, the transmission of knowledge, which has defined a lot of the structural power systems in the university and is led by professors, has today become pretty mundane.

The third part is jobs. AI is already re-shaping the job market and universities need to think more seriously about what they are preparing students for. This will require big changes in what is taught in universities. Today, those who control universities also control a large share of investment in knowledge, and they have little incentive to change. As long as demand remains strong – students still queue up for admission to Oxford, Cambridge, Cornell, and Harvard – there's no real sense of urgency. It’s the classic “frog in boiling water” problem: professors hear talk of disruption, but their day-to-day reality still feels stable.

The real pressure is falling on weaker institutions. Many lower-tier universities are already in crisis. They lack the resources to adapt, and by the time they react, it may be too late. The bigger question is, what will prompt change amongst the strongest institutions, which still have the capacity to act but no immediate need? I'm not convinced that kind of change will come easily in the West.

Gavin Allen: Do you think those in emerging markets will leapfrog?

Prof. Soumitra Dutta: There’s certainly more willingness, and China is a good example. Their universities were in a pretty sorry state years ago; today, many are world-leading. They had the ability and the desire to change. They are bolder at innovating. For example, Harbin now offers a PhD earned not by writing a thesis, but by creating a product. It is more designed towards promoting scientific innovation and building companies based on products. That’s not the only solution, but in the West, that idea would hit strong headwinds from the current faculty. They’d give you 10 good arguments for why nothing should change.

Gavin Allen: So how should universities push for change?

Prof. Soumitra Dutta: We need a counter-argument from a visionary leader within the system. Arizona State University’s president is a good example. He has invested heavily in technology-based learning, while funneling profits from that initiative into hiring top research faculty members. Without such decisive leadership, change stalls because the system remains designed to suit the average professor, rather than their students.

Gavin Allen: Do we have to change how we define talent in this AI era?

Prof. Soumitra Dutta: I'm struggling with that. If AI can do a lot of what human beings do, then it raises the fundamental question, “What is talent?” And then, “What is the role of universities in cultivating talent?”

Humans do three things well. We identify patterns – in words and data. We create new patterns, reorganizing data to form new concepts and ideas. And we build relationships with people.

AI can already identify patterns better than we can, and increasingly, it can create new ones faster. Even the third pillar, which people say is uniquely human – well, machines are now getting people to fall in love with them, and some patients prefer machine-based counseling over humans.

If you assume AI will become as good as humans on all three dimensions, and will put itself in a self-improvement loop, then what value will the majority of people add? If we fail to find good answers for this, then we start raising fundamental questions about preserving human dignity.

Having said this, AI is also creating all kinds of new opportunities which require new skills. For example, while software coding is getting automated, thousands of new jobs in integrating agentic AI within corporate IT systems are also appearing. But this requires universities to change and even more faculty within universities to also change what they teach. Universities need to prepare for constant change, to adapt and be resilient. But they’re often more focused on maintaining the status quo.

Gavin Allen: How can they do that?

Prof. Soumitra Dutta: We really should be forcing people to use AI as much as possible from as early an age as possible. Some people want to remove technology from the classroom, but we need to learn to adapt, so we should actually immerse ourselves in it, because this AI-intensive world is not going away. We need to embrace a much more experimental approach.

Often in organizations, change is less about vision than about courage, and being willing to push all the levers and fight all the battles.

Most universities evaluate faculty for tenure, which is a permanent position, on three criteria: research, teaching, and service to the institution. At Cornell Tech, we pushed hard to add a fourth criterion, called impact. It’s amazing it wasn't there before, but it's not even there today in most universities. Most are happy just publishing in prestigious journals very few people read. That's how they’re evaluated, and that’s what the whole system rewards – not, as it should be: Have you actually created X number of jobs, or helped improve X lives in a meaningful, tangible manner?

It’s so difficult to get faculty to change. Education is really about creating a brand associated with a university and its professors. But what if you could take that brand and create content at very low cost, then market and distribute it more widely? With avatars, the same professor could become 50 professors, multiplied for different audiences, cultures, and languages. And that professor will understand your learning journey, just like Amazon understands your shopping journey. We should be building a system that allows you to do that.

Gavin Allen: So, you, Professor Dutta, would be here, teaching this group – and simultaneously teaching other groups in multiple other locations? The saying “I can't be in two places at once” will no longer be true?

Prof. Soumitra Dutta: Exactly. On the knowledge side, technology can pretty much replicate what needs to be done. The social side is harder because it requires a physical presence. A colleague of mine was president at a Korean University that went online briefly during Covid. When the school re-opened, students said they wanted to keep taking classes online and have the university organize more in-person social and job-related gatherings for them.

That’s what makes the Oxford college system so brilliant. It’s the social functions and gatherings, the social development that Oxford provides, rather than just the professors. Maybe in the future, universities will focus more on the social development side, exposing students to different careers and other kinds of experience.

Universities must evolve into smaller, research-focused hubs that function more like start-up incubators than traditional schools. By integrating AI tools and elite students with a “Y Combinator” model, we can move beyond mere lecturing to focus on incubating ideas and launching companies. At Oxford, I envisioned marrying the college structure with this entrepreneurial training to create a pipeline for jobs and social impact. Since existing institutions are often too resistant to change, we must create a new breed of what might be called “opposite colleges” – schools whose mission is centered on creation, rather than just contemplation, where the goal isn't just to graduate, but to innovate.