To get things started: this entire project is about AI's disruption of entry-level employment and what universities should do in response. When I talk about that subject, what are the first things that come to mind, and your general opinions about it?

I think you're correct in that the pyramid is inverting. Before, we bloated up the entry-level layers, and they were at the bottom of the pyramid. But I see a pyramid inverting because, at the end of the day, what a lot of people still require is judgment and critical thinking. And while AI does that, in this part of the world we also need to realize that context matters. And if we don't have the trained models that work properly for the decisions and the problems that we have in this market, it will not be as useful as it would be in a different market. So while we know that yes, AI is disrupting, AI is here, entry-level positions are threatened.

I would still say with some level of caution in the Global South, especially sub-Saharan Africa, because when you think about even affordability and access to a lot of these AI tools and models, you find out that not many people are using the paid versions. They're still all using the free versions. Not many organizations are deploying beyond Copilot that comes bundled, for example, with your Microsoft licenses. Not many people are still going on to say, how are we using AI? A lot of this part of the market is nascent. Some of it would be pulled down from within the multinational corporations in terms of bringing on some of these tools and utilities.

But in terms of the main heart of Nigeria, the heart of Africa, the millions of small and medium-sized organizations, AI's impact of disruption is still further away. But that doesn't mean that we don't need to look at entry-level positions and start the preparedness for what could come. And I think entry-level positions in this part of the world are still at risk because what you find is that you have graduates who have come out of colleges with credentials. But then again, the size of the formal versus the informal sectors is still very wide, and there's a real missing middle there.

So how many of the small and medium enterprises that are more prevalent can really absorb graduates and pay them for the level of education they've attained? So what you find is that there's still an oversupply, or a demand for professional organizations or corporates rather than the smaller and medium and informal organizations, when it comes to these, especially graduate entry-level positions. So for us as universities, one of the things we need to focus on is to take away all the tech and all the AI and different things: I think people still want people who have some form of critical thinking skills.

You want people who have judgment, critical thinking, can make a decision, can analyze data, and can really know what to do and navigate some of these things. So with or without the technology, you still need that human in the loop that has the worldview of what we're trying to achieve and will bring that perspective into the conversations. Again, because a lot of the businesses we do are still somewhat social or relational. You need people who can communicate, who can engage, so we can't leave everything to the machine.

So for us at the business school and the university, what we pronounce is that we are teaching leaders and preparing leaders for the future market. And those leaders have certain ethos or characteristics. One is ethical thinking and reasoning. Another is a global perspective to understanding what's going on outside of the market and the context in which you're living.

Another is communication and critical thinking. And yes, I think those are leadership skills. So whether AI is here or not, we still believe in the formation of the person, and that person needs to have some of those characteristics, if not all, to ensure that they can be appealing to the market.

You said something interesting there. You said take away all the tech and AI, so that people can have critical thinking and people skills and that sort of perspective. Do you believe that using AI is harmful to those aspects?

I don't think using AI is harmful to those aspects per se. It's about, you still need to show up. You can't show up with your AI. And so even if you use your AI to prepare, you still need to show up. You need to be able to frame a logical argument. You need to be able to put that position in a meeting or amongst colleagues. So for me it's the complementary use or the augmented use of AI: not that we get rid of people and replace them with AI, but you bring people who can use the AI and bring the AI and their judgment and their thinking into the conversation.

If you take away Gen AI from this whole conversation, I don't think we'll be having a conversation like this today without Gen AI being in the market. I think Gen AI just propelled and brought everything to the fore. So if you think about it, we've had AI since the 1960s. But only the geeks and the techie people knew what AI was and the different forms of AI. But with Gen AI, now it has come to the mainstream.

And now, when you think about the mainstream, yes, we can all use Gen AI, but for me, the business case for AI still lies with what we do. How does it really enhance our businesses? Because a lot of people who are using Gen AI are using it individually. So when you go to a corporation, how much AI are they using to transform their business? Either using it for forecasting and predictive analytics, or using it for their workflows and their processes, or to analyze the vast amounts of data, you find out that that's still very nascent. Because the investments they even need to make to be able to do those kinds of things are not really there yet.

There's a raging debate on what AI is going to do in reducing people in the workforce; it's one of the reasons we launched this project. However, in your AI in Nigeria whitepaper, you found that only 13% of sub-Saharan organizations reported AI-driven workforce reductions, and that 88% of Nigerian workers believe AI tools will help them develop new skills, creativity, and better work quality. This goes against what some industry leaders and financial institutions have been saying. Can you tell us more about why you found opposing results?

The thing about it is, AI does not work in a cocoon, in a bubble. It's within an ecosystem, and the nature of the ecosystem would really determine how AI is going to be used. So when we think about work reduction and things like that, a lot of the corporations and small and medium enterprises here haven't even gone through a digital transformation. So use of AI and use of tech is still very nascent. Basically, the bottom line is that it's cheaper, easier, faster to hire a human than to go through a digital transformation process. So we're still at that crossroads. So again, it's not really job displacement that we're seeing; it's more skill augmentation in terms of the workers needing new skills.

I think what AI is doing, especially in the formal sectors, is that the nature of work is going to change. And the work tools that we're using are also going to change within a work environment. But again, those will still take a lot of time to really ingrain themselves into building out new work systems. So when you think about African institutions, for example, a lot of our government institutions are still very manual, and this is where the opportunity for AI comes in. But that opportunity also needs to be well managed, because one of the biggest challenges and risks is the lack of digital tools and digital transformation.

So when you think about the whole government system, you're not just looking at AI within one person, but how can the entire institution use AI? And that really is going to mean a really different organizational design. How many people are ready for that kind of impact and disruption and change? So you realize that you find one or two people using AI personally and privately. But in terms of that whole transition and transformational shift, that's going to take much more effort, much more systems-level thinking in terms of the work we're doing, the work processes, and then from the government perspective, because some of these tools or some of these systems are also enshrined in laws.

We have to go back to the legislature and say, what does this mean in the real sense? And I'll give you an example. When Nigeria wanted to launch the eNaira, the central bank digital currency, I think that was in 2021, one of the challenges was that the Central Bank Act specifically talked about minting currency, which is notes and coins. But a digital currency is no physical notes and coins. So they needed special dispensation from the president. Otherwise you'd be violating the law. So I think that a lot of work in this part needs to be done regarding how we use AI, and use AI to really refresh and modernize a lot of our systems, practices and processes.

But that takes a lot of work, a lot of investment, and again a lot of reorientation and upskilling of people. Because for me, it's not about job displacement. The jobs are going to be pushed somewhere else, because you're going to need more testers, you're going to need more assessors of different types of tools, or when you're training your models. So it's really just changing the nature of the work that we're doing.

In IWA Magazine in March, you said that AI is not just a digital tool but the architectural foundation for how businesses create and deliver value. If AI is the foundation and not just the tool, what does that mean for LBS entry-level graduates? What do they need to do on day one that they did not need to do five years ago?

Well, on day one, they need to challenge everything with the tools they have access to now. It's the same as when the Internet started: before you could Google or Netscape or anything, you had only what you had access to from a physical library of physical sources. But now the first thing you need to do is basically validate what you're doing. And I think for an MBA graduate or any student of Pan-Atlantic University, it's really assuring that you're working with global practices and you understand what global work looks like.

So we expect you to bring in all those capabilities into your work system, your work process, and to be able to use them not necessarily to dumb down your own argument, but to validate what you're doing. One of the things I always say is that you write whatever you want to write, and it's almost like you write an email and you have Grammarly installed, and Grammarly helps you refine it to make it softer. But the core idea still came from you. It's really about still having your core ideas.

But then again, you now have an opportunity to validate your core ideas in a bigger space, in a bigger world, and then refine them. And for me one of the best ways of using Gen AI is really in a conversational manner. Oh, what would this look like? What would that look like? Oh, but have you thought about this? Almost like you're having a conversation with a colleague, rather than having a colleague in the room. And then you bring all these things in and you bring it to your team meeting and you actually disclose that, by the way, this is what my Gen AI tool says, what do you think? And we continue using that to ideate and create.

So for us, it's the way we use it and let students realize that, again, a lot of the discourse on university campuses is really about exams and assessments. And how students are going to cheat, and how we're going to need to change how we assess. But I think right now it's really not only about assessments, because you now ask yourself, what do assessments do? They validate learning. Sometimes you might need to realize that what AI is doing is not to invalidate or to help them cheat, but really to expose their thinking. And it's now for us as faculty as well, to enlarge our own thinking sets rather than staying within a narrow space, and also help the students challenge and use the tools to creatively continue to think ahead.

But one of the biggest obstacles that we are seeing is that there's also resistance from faculty, especially when it comes to student use. We want to use it when we're doing it for our own work, but when it comes to students' use, some of us are used to, oh, let me just take the easy way out. I'll set the same exam, AI can answer, and things like that. So we're all human beings. We all look for the shortest path. And if AI is going to give us that shortest path, we have to go back there. But for me, it also goes back to, what are the learning goals of the programs or the courses we're teaching?

And how do we, with or without AI, ensure that we're instilling and enforcing and emphasizing those learning goals? Because right now people don't want single technical experts in a domain area. People want people that are more interdisciplinary and can think beyond or see beyond their own domains, and give life to what happens outside domains. And AI helps you do that. If I were a chief executive, what would I do? If you were a CFO, what would you do? You can build all these kinds of roles and play them without really leaving your doorstep.

That is a great answer. How does that look practically at LBS right now? How are you helping the students challenge the current way and test tools? How have you adapted the entrepreneurial classroom to fit the modern AI world?

Right now I'll give you certain examples. And I'll tell you that right now my focus is first of all on the faculty. Because we can't do anything if the professors are not also using these tools to challenge the status quo. So we've been working on the faculty to basically onboard them. And don't forget that faculty, because of their own domains and the traditional ways of doing things, are also set in their ways. So change management is one of the biggest things that we're working on now.

The second is having general AI education across the faculty. Because when you can think about it from a productivity perspective and use it from a productivity perspective, what we've observed is that you're more inclined to use it in other settings. So when you can see how it works from a productivity perspective, and that's what we're also encouraging, let's let them be able to use it for their own work and productivity.

Then the third layer is: how do we use it with students? Some of the ways we use it with students are, for example, go and do the search. Use AI, brainstorm with AI, and then come back into the class and form your own argument. So you use the AI for your brainstorming. But we want the original thinking to still come from you, and for you to be able to communicate and express what you're thinking, what you're saying, with or without the AI.

So we're doing a lot more of verbal and oral communication, because a lot of our classes are participant-centered. So even when you use AI to find an answer, and I ask you, and so what? That thinking and that logic still needs to come from you. So why do you believe that AI's answer is the right answer? You have to defend and justify.

Some other colleagues in finance use it for their financial analysis. Some others are telling them to go and do some vibe coding and see what they come up with. So for me, we're happy to give them the challenges, because in giving them the challenges, you're also giving them the exposure, and you're also helping them see where AI is vulnerable, or where AI does not quite work. Because sometimes telling is one thing, experiencing is another. So they can now begin to see where AI starts to hallucinate.

And in a product innovation class that we teach, we ask them, using the Internet and using Gemini, because Gemini has vaster access to the Internet, to analyze your company. What types of products would your company need to develop to solve financial inclusion? So it helps them in doing that research from public sources.

Your paper called for AI curriculum from early school age and partnerships with professional bodies. Sitting where you sit as Dean, where are business schools too slow to change at the moment? And what is the one single change you would want LBS and every other business school to make in the next 12 months, the first step, so to say?

One of the things is we need to start building. AI has become so ingrained, and we need to start working through and helping people understand when to use and when not to use, and what it can do. So that education and literacy needs to start from the beginning.

From an LBS perspective, business schools are really late to the party in my sense, because we've been used to: how do we write case studies? How do we develop intellectual contributions? How do we do our research? Without really asking ourselves, if you think about that whole workflow, how do we use AI even to enhance that workflow?

And I think that it's not just business schools, but the whole scholarship and academy domain of management has to rethink. So if you think about it, why do we even need to publish journal articles? It might take me three years to publish an Academy of Management Review journal article when, by the time you're publishing it, the content you're publishing is already stale.

So how do we start to think about using AI to put more things out there, but again, in a responsible manner, and then it's the filtering and all the other ethical guardrails that we need to put in place.

For me, what would I like to do? What I'm still trying to do is, I've been teaching at the business school for 23 years, and in that period I have a lot of knowledge, class preparation, and material amassed. One of the things I want to do personally is to build my own bot. Because I find out that information now is democratized. Why do I need, as a professor, to hoard information? Let me give you an example: if I continue hoarding information, how useful is my voice in the conversation? If we hadn't done some of the work we had done with the whitepaper and some of the other things that we had written and said, you would never have found me.

And so when you think about what we're doing here, I think what we're trying to say is, knowledge is available in different domains. And especially for us as professors in the Global South, how do we start to use technology tools, AI, to even promote our voices, our opinions? Because there's a lot of entrepreneurial knowledge and indigenous entrepreneurship that sometimes you don't know about.

How do we help people who might be moving from, let's say, the Global North to a sub-Saharan African country? How do they deal with power cuts? How do they deal with traffic? How do they deal with rain? Because those things are not codified anywhere. But how do we use AI and start to build these knowledge systems?

And for me, how do we start to use AI to influence business school curriculum in the Global North? Because we know what we teach and why we teach it, based on our context. But a lot of business schools in the Global North have advanced in their problems and their solution sets, and that doesn't mean they would not need to understand what global work looks like. Global work is no longer just about the Global South. It's also the rest of the world, especially as we're working in global teams. So how do we start to take that knowledge and that education even from south to north, rather than the other way around?

So I think AI has a lot of promise in helping us codify and share, especially as you think about vast amounts of information. Maybe vast amounts in my world, which is big, 23 years of work, but then again, it's a small language model. So how do we help people begin to put their knowledge out there? Because if you don't put it out there, somebody else will. And then you're really going to be an unknown personality for the rest of your life here.

AI also has made it easier. I'll give you an example. A colleague of mine is a good professor of accounting, and I said to him three years ago, why don't you start doing YouTube channels, and you can monetize them teaching accounting, because it's a skill any business person needs in terms of reading financial statements. But today he doesn't even need to do that. AI can do it with his knowledge. If he gives an AI the tools and the notes that he needs to create, AI can create a class. So it's just putting everything on steroids.

We did have one of our interviewees whose university actually made a digital version of him, and his digital version has taught around 30,000 students. It's what you're talking about.

Exactly. And I think these digital twins are important, because there's only so much you can do as a human. And if my digital twin can be teaching someone in China when I'm sleeping, hey, it works.

Through the Microsoft AI program, LBS chose to train 99 senior public-sector leaders, members of the National Assembly, and executives from ministries. What was the strategic decision that led you to train the top layer?

One of the things we're trying to do is: how do we start to bring this AI mindset and digital mindset into public-sector leaders? And it was one of the things that, let's start to test and do. Because at the end of the day, decision-making starts with the top. If the top person does not understand what you're talking about, the natural inclination is to shut it down. Because I'm not going to promote what I don't know or what I don't understand. So what we're trying to say is, how do we start to even get them thinking around AI?

How do we get them not just thinking about AI for the sake of it, but looking at their roles and trying to identify what problems and use cases AI can address in my own job context? Because the knowledge is there, but knowledge applied is really what we're helping them think about. Because when you think about that, then that helps them begin to appreciate what AI can do, rather than the narrative of AI is coming to take your job.

So the context is, how do we help them understand it? How do we help them get comfortable, at least introduce them to it, that it's not about taking them or replacing them, but really adding value to them. And then based on that, we can now start to get them to think about it from their perspectives and continue to work. So that's what we've been doing with that, and we had 100, and based on that hundred, we want to continue to do that and expand it. Because we believe that once you can get them engaged and involved, it changes how they think about AI, it changes how they work with technology, and it continues to embrace them. So when technology decisions or discussions are being made, they don't shut it down.

I hear you, and it makes total sense. It seems similar to what you're doing in your own university: teach the faculty first, and then the students.

Yes. Because I think it's two ways. There's a top-down and a bottom-up approach, and I think we need to address it both ways. But the faculty are the ones that we want to change a bit more. So how do we ensure that they can change and they can adapt, and support that adaptation? Otherwise you'd have irresponsible use of AI. And we don't want a situation where faculty don't know, and then the students are the ones challenging the nascency and currency of their learning because of faculty's knowledge sets.

In 10 years, what does the first job for a 22-year-old graduate actually look like? And what is the one thing about it that universities are not preparing them for today?

That's a real crystal ball. I think in 10 years the 22-year-old is going to be working with training models, building and training models. Because whether we like it or not, we're going to have to continue to generate that content that is relevant for the models. And the models would be quite advanced, and 10 years is a long time. The models would have taken over a lot of what we're already doing. But I think in 10 years it's really going to be training models. There might not even be that entry-level position per se, in the real sense of it.

Because if we take all our workflows and build agents, it's really going to be different agents working through our system. So what are these people going to be doing? I think they'll be validators. They'll be validators of what is going right, almost like traffic wardens. They'll be able to pass what's good and stop what's bad, because you still need that ability to process very quickly and assess what's going on without really stopping. Almost like air traffic control is another analogy we can use in that sense.

So how do we start to prepare them for these kinds of things? I think two ways. And in a way it's not really what universities are doing, but how these young people are also engaging. A lot of them live in a gaming world. And that gaming world already gives them the physical abilities to be able to use these things. The other thing is how do we connect that with the real world, and the real world of work and the world of business and commerce. So for me that is going to be about bringing in more experiential learning, and getting them to test and learn and test and learn.

Another thing we're doing at the university is, we have an art museum on campus. So we also get students to spend time in the art museum, in courses with object-based learning workshops. And those object-based learning workshops help them broaden their thinking beyond just their core curriculum. So if it's a science or a computer science student who is just thinking about coding and programming, we think about object-based learning: take them into a different world, into a different context, and see how they can draw the relationships and the analogies to that. But it's still really about the brain's ability and capacity to expand and grow and understand different things and new things.

I think that's a wrap. I feel like I've gotten everything that I wanted to ask you. Are there any closing remarks you'd like to add before I shut off the recording?

One thing I would add is, as you're thinking about this from a demand perspective, also let's think about the supply side. Because a lot of the time we're focusing on the students as they come out. But students will only be as good as the professors that develop them.

So how do we also ensure that we're building those? And again, you mentioned something in your introduction, Alfred, about the role of the university. If the role of the university is changing, the type of universities we're building today is changing in another 10 years.

What do we need the university to look like? How would it look? What would it feel like? Do we even need four-year programs anymore? All those things are being challenged, and for us in the academic sector, we've always had that comfort of sustainability, job security and things like that. Even we are now having to think that in another 10 years, what is the role of the university, what's the role of business schools, and how do we start to reorient ourselves around it? Because technically, for the last hundred or so years, academic institutions have been protected from any types of shocks.

So now we're vulnerable, and how do we start to re-engage? You find out that some people will say, you know what, I'm just going to retire, because the transformation is within. Then how do we start to build these AI-first universities? And that means we have a lot of content and knowledge. But again, like everybody else, we haven't focused on making it digitally available, because we've talked about IP, we've talked about all these things. But I think the transformation and the transition of what a university will look like in this new era has to start now.

And unfortunately, when you think about some colleagues I've been talking to that have millions of dollars of budgets to do this transformation, we don't have that in sub-Saharan Africa. So how do we fund it? How do we fund having access to GPUs and compute? How do we fund having access to training models and building the systems from the ground up? So that's, for me, my fear and the challenge and the risk. So again, you find out that the digital divide is going to continue expanding, because we're resource constrained.

*This interview has been lightly edited for clarity and readability.