The human factor in AI with Prof. Sherif Kamel

Sherif Kamel: AI is not software; AI is an environment. It's a new team member in the football team. How can we use it? Is he or she better on the left side of the pitch or on the right side of the pitch? And should we use it throughout, or better to go back in the second half. I mean, I'm just doing an analogy. 

Patrick FitzPatrick: Yeah. 

Sherif Kamel: I look at AI more as augmented intelligence than artificial intelligence. I see this as human intelligence — I see it more as HI. 

——— 

Patrick FitzPatrick: The noisiest conversations about AI right now are happening thousands of kilometers from here, in San Francisco, Washington, New York, and in Abu Dhabi, Riyadh, and Doha. Globally, you've got the accelerationists who think AI is going to cure everything, the doomers who say AI is going to destroy half of all entry-level white-collar jobs in less than five years, the bubble guys who think the whole thing is just a very expensive magic trick.

At the same time, they're imposing export controls that block the rest of us from using the best chips and the best models. Every one of those arguments is a rich-country argument. You impose export controls when you own the models and when you own the chips. You debate universal basic income when you're the ones capturing the profit. You talk about an AI sovereign wealth fund when you assume the wealth is coming to you. And none of that is about a place like Egypt. We're not even non-player characters in this debate.

Not far from here, the Gulf bought its way to the table. Abu Dhabi, Riyadh, and Doha are pouring hundreds of billions of dollars of sovereign money into AI. Egypt's plan? Well, we're going to train 30,000 specialists and half a million outsourcing workers to serve everyone else's clients. The Americans are arguing about winnings, the Gulf is buying a seat at the table, and we're going to sell labor by the hour.

I think it's time we started having a conversation about what AI really means for Egypt—not is it good or bad, but what it actually means for the economy, for our companies, and for our kids.

My guest today is somebody who has watched this movie three times already. He's had a seat at the table as Egypt adopted the personal computer, the internet, and the mobile internet. He's done it inside government, from inside the university, and sitting on the boards of businesses and institutions that are grappling with it all.

Sherif Kamel is the dean of the Onsi Sawiris School of Business at the American University in Cairo. He sits on the board of the Central Bank of Egypt, and he's served on boards including CIB and the Egyptian American Enterprise Fund. Sherif has a way of making the terrifying feel kind of just like another Tuesday, and maybe that's the most useful thing anyone can offer us right now. Maybe, just maybe, this doesn't have to be so frightening after all.

I'm not going to tell you guys what to think. I'll let Sherif walk you through it, and then you make up your own mind.

And that's the conversation.

Patrick FitzPatrick: Sherif, welcome to Making It.

Sherif Kamel: Thank you.

Patrick FitzPatrick: So, listen, man, before we get into the AI of it all, I want to go back in time. You saw the internet arrive in Egypt before pretty much anybody else. What was the first time that you understood what this thing called the internet was, and what difference it could make to a country like Egypt?

Sherif Kamel: Well, that's a big question and I have to go back almost 45 years.

Patrick FitzPatrick: We're showing our age today.

Sherif Kamel: Exactly. The internet came in a very sudden way. We started to experience it at the university, but also shortly after I graduated, I started using it at work. I used to work for the government, and I actually used to work for the entity that introduced the internet to Egypt.

Patrick FitzPatrick: This was nineteen ninety what?

Sherif Kamel: This was 1991. It was working, and at the same time, I was doing my PhD. The very first time I saw an email was actually in London at LSE. A colleague of mine showed me how you can send a message from a screen to the other, and for me, [it was] "What is going on?" and the rest is history.

So, you look at the internet 45 years later, and it's gone a long way. But it's not just the internet; it's the acceleration of digital transformation, and the internet was just one element of it. You and I remember those days when you finish work, you go home, you're done. You now live in an inter-connected enterprise. Now you live in an interconnected enterprise. You're at home, you're in the car, you're anywhere — you're connected. That creates opportunities, but it also creates distraction. 

Patrick FitzPatrick: Yeah

Sherif Kamel: And it’s up to us. Again, you’re going back to the most important element in the infrastructure, which is the people factor. In my view, the most important asset in society. They decide how to strike that balance: how to use technology, when to use technology, and for what.

Patrick FitzPatrick: When did you know that it was more than watching someone send a few characters on one sort of green-on-green phosphorescent screen to another? When did you understand that this was something that could help transform an economy?

Sherif Kamel: In the early times, when we realized, everybody, not just me, that the difference of time and distance is no more. So, you don't have to be in the same room; you could be miles and miles away and.. 

Patrick FitzPatrick: You’re talking, you’re transacting. 

Sherif Kamel: Absolutely. And it's not just increasing productivity, but also becoming more efficient. So you look at the acceleration from the, I would say the early '80s when some of us started buying personal computers at home, and then the mid-to-late '90s with mobility, and so on and so forth. The acceleration gets more and more and more. Maybe the most recent chapter is the GenAI. But people think GenAI or AI in general started in November 2022. Absolutely not.

Patrick FitzPatrick: It wasn’t ChatGPT that started it all.

Sherif Kamel: In the '90s, we used to call it the expert systems. That was still based on some form of AI.

Patrick FitzPatrick: Is this just the next wave of a continuous technology, or is there a real discontinuity here? Is there really something new here with AI as we know it today?

Sherif Kamel: AI as we know it today, I think it's both an acceleration, but a shift, because we're introducing a new member of the team. It's a new team member that we need to know how to use it. So I see that as the next phase in the acceleration of digital transformation. We probably would have called AI science fiction 20 or 25 years ago, even with the versions that we had at the time. Now it's no more. It’s not science fiction. It's a platform that is out there. 

Again with the simplicity of using it, there’s so many questions that are being raised about security, privacy, bias, inclusion, inequality, divides. Does anybody have all the answers? I don't think so. We're still learning as we move because we're training a machine to think very fast. Should we rely on it 100%? Absolutely not.

Patrick FitzPatrick: How are you using it at the university today?

Sherif Kamel: So I look at universities around the world, and not necessarily business schools, because I want to think when it comes to AI, that it has to be multi/inter disciplinary. Every single university under the sun is thinking very fast, trying to put resources on how to deal with this. And I would rank them in three categories. Some of them who came up with 100 pages of do's and don'ts.

Patrick FitzPatrick: Yep

Sherif Kamel: I still think that's not the answer, because the way we saw AI grow is more organic. Even faculty, they don't use it the same way. And there are universities who came up with a statement. It’s an open-ended statement that this is [AI] an opportunity. This is a platform that we should look at how to use it, when to use it, how it’s impacting the pedagogical approaches on campuses and so on and so forth. 

And there is a third category, which I think are in danger. The ones who [say] “okay, let’s wait and see.” Wait for what? And see what? I mean this is.. 

Patrick FitzPatrick: You think they're in the worst position. Even worse than the people with the 100-page rulebook of do's and don'ts?

Sherif Kamel: Absolutely.

Patrick FitzPatrick: Why are they in the worst place?

Sherif Kamel: Because they are losing out. Like the machines that are being trained to be smarter. We are getting smarter at using that. So the more we are leaving ourselves behind, the gap will widen and the technology doesn’t stop. But I also want to be very conscious by saying: there is a hype and there is a real impact. I mean, most of what we’re reading is the hype. 

Patrick FitzPatrick: Yeah, the world is kind of divided right now into doomers, accelerationists. People who are saying this thing is so powerful, that’s going to lead us to utopia or disaster, on one hand. You’ve got a category of people on the other who look at it like Ed Zitron and say “this stuff is, it’s a magic show.. It’s a really overfunded, subsidized magic show.” How do we look at this type of debate and figure out where we sit in Egypt?

Sherif Kamel: Very interesting question. I would start with four categories that we need to look at. And it's very important that we look at these categories in the order I'm going to share them: people, data/information, governance, and then infrastructure. You would realize that most people, when they talk about AI, they start with the infrastructure. Yeah, they start with the —

Patrick FitzPatrick: Yeah, how many gigawatts or megawatts of compute. "Am I  building through what pipes?"

Sherif Kamel: — Am I saying that infrastructure is not important? Absolutely not. But in the order that we should prep ourselves, is the talent, the data that we're using — I mean, it's all about data — 

Patrick FitzPatrick: Yep

Sherif Kamel: — and then the governance model, the policies, the procedures, the workflows, and then finally the infrastructure.

Patrick FitzPatrick: And that's at a society level?

Sherif Kamel: That's at the society level, that's at the organizational level, that's even at the individual level. So give me the most up-to-date technology and I'm not knowing how to use it—granted, some will say, "but it's very easy to use it." Yeah, what level of use? And here comes the adoption. I mean, we heard a lot about prompt engineering, and, and, and...

Patrick FitzPatrick: That had a thing, that had a moment.

Sherif Kamel: That had a moment. But again, when I think of prompt engineering, I'm also only thinking of traditional AI, not agentic AI. And there is a sea of difference between the two.

Patrick FitzPatrick: Explain that to our listeners.

Sherif Kamel: So in a very simple format, traditionally, I ask a question, you get an answer; stop.

Patrick FitzPatrick: Google with a bit more personality.

Patrick FitzPatrick: Mhmm

Sherif Kamel: It's dependent. There is no cycle, there is no chain. But then agentic AI is autonomous. It can do the job. You structure it in a way where, uhh, from start to finish, there is no human interaction.

Patrick FitzPatrick: So I can feed you a PDF or an Excel with a few years of financial data and say, "Okay, go build me a model, stress test it this way, validate your output, make sure that you've made no mistakes, then bring me back a final product," and the machine goes and whirs and clicks and builds.

Sherif Kamel: Absolutely. And it's going to go and maneuver and think of creative ways to get to the answer. So it does the complete task. But then what most people use at this point is traditional AI, which is asking a question, getting an answer. When it comes to the agentic, it's very judgmental. And I think here is the crux of the matter: being humanistic, we need to really be careful because it's a machine that is judging. So when they went to a path and they did not complete the task, they started working in different paths until they completed the task. We need to make sure that this judgment is what we want.

Patrick FitzPatrick: And if this thing is working for 45 minutes or 45 hours autonomously, are you really going to go back and look at its chain of logic and read through pages of transcript to 

Sherif Kamel: Good question

Patrick FitzPatrick: see what trade-offs it made or what decisions it made, if future models even make that available to you?

Sherif Kamel: Tech companies will tell you, "Listen, this is machine learning fast." I get that, but I still think the human factor—how to critically think, how to factor in elements that might not be, sort of,  obvious — will remain the most important factor in reaching the optimal solution that we're looking for.

Patrick FitzPatrick: Because we can do something that it can't, because we have something that it doesn't.

Sherif Kamel: Because we can think differently, we can value some elements in the equation that might on paper be not really important. I think computers at this point cannot do that. And allow me here just to go back to the very definition of AI. From the premise that the human factor remains crucial to development and the acceleration of digital transformation, I look at AI more of an augmented intelligence than artificial intelligence. I see this as human intelligence... I see it more as HI. At the end of the day, who's creating the model?

Patrick FitzPatrick: I think the worry is that the LLMs start training the LLMs and creating the LLMs —

Sherif Kamel: Fine!

Patrick FitzPatrick: — but today it's humans. It's humans. 

Sherif Kamel: It's humans!

Patrick FitzPatrick: It's humans. 

Sherif Kamel: Can today someone go through the cycle and come up with the results and so on, and just submit it without looking at it? 

Patrick FitzPatrick: Some will and they’re — 

Sherif Kamel: Very wrong.

Patrick FitzPatrick: They're not thinking... 

Sherif Kamel: Very wrong.

Patrick FitzPatrick: Yeah.

Sherif Kamel: I'll just share an example. Let everybody pick one of the AI application tools that they are using and just ask about themselves. "Who is Sherif Kamel?" for example. He'll end up with some false information. 

Patrick FitzPatrick: Yeah

Sherif Kamel: Some will argue, "Yes, of course, today. But next year, maybe in six months, in 18 months, you ask about yourself, you'll get all the information right." Still, you cannot just use that outcome without reviewing it, making sure that the flow, the judgment is what exactly you're looking for.

Patrick FitzPatrick: What do you use it for personally? And who is your AI of choice?

Sherif Kamel: So my AI of choice, after trying a few ones, is Claude.

Patrick FitzPatrick: Okay.

Sherif Kamel: I cannot claim until today that I use it heavily in my work, but at the personal level, fact-checking is important. So what do I do? Sometimes I cross-check between multiple AI tools, and guess what? I get different —

Patrick FitzPatrick: — different answers.

Sherif Kamel: Absolutely. Absolutely.

Patrick FitzPatrick: We both like Claude. We're increasingly using Claude across the business, from the newsroom to finance. Anthropic, the model maker for Claude, recently had to pull access to Fable 5 and Mythos 5. The US government decided no foreign national, including Anthropic's own staff, should have access, and Anthropic's response was to yank access. Looking back to the late '90s, early 2000s, I don't ever remember thinking that somebody could turn off a fundamental technology for us. How concerned should we be about this?

Sherif Kamel: It's a different world. I used to tell my students, "Watch the movies, because the technology we see in the movies, once we saw it in the theater, that tells you that it's already obsolete."

Patrick FitzPatrick: Because somebody is already trying to build it.

Sherif Kamel: Of course. I mean, the internet itself is a very simple example of that. We knew it, like, what, 30 years after it was in place?

Patrick FitzPatrick: The defense... DARPA.

Sherif Kamel: DARPA.

Patrick FitzPatrick: Exactly.

Sherif Kamel: Exactly. So the more we know, the more we realize how much we don't know. Over the last 30 years, things have changed. Today, countries are looking more for sovereign systems. They want to keep it indoors. Unheard of. I mean, we've always thought that...

Patrick FitzPatrick: "Information wants to be free," they kept telling us.

Sherif Kamel: Absolutely. And the democratization of information. I think it’s still there, but up to a certain level. There is a threshold that, given the competition that is taking place — not just the national interests when it comes to defense and the like, but also the business interests...

Patrick FitzPatrick: Whose companies get to make those profits.

Sherif Kamel: Absolutely. Absolutely.

Patrick FitzPatrick: If you were still in government today and you were advising cabinet in terms of how to view that, how should they view that?

Sherif Kamel: I think we have to have a sort of balanced approach. I would definitely — and I think that's already happening—look into a sovereign sort of platform that is local, that would serve the interests of Egypt through models that look at talent, governance, infrastructure, data, policies, and at the same time, not really build walls around ourselves. So it's important to evolve as the technology and the approaches that are used worldwide evolve with it. We sit in a location I think is extremely strategic, extremely important. And if I go back to my argument that talent will remain core, I think we have quite a lot of talent that we can train to be game changers in that space. 

Patrick FitzPatrick: Yeah.

Sherif Kamel: The next battlefield—and I'm sorry to say that word—is all about talent.

Patrick FitzPatrick: When we come back, we're going to get into the question that I called Sherif about in the first place. Egypt's AI plan is to train people and then watch them leave, like we've done for generations with doctors and engineers and HR professionals. Sherif built our last national IT strategy, and I want to know: what would he do instead?

———

Patrick FitzPatrick: You know, you go to Abu Dhabi and they're talking about building a campus with five gigawatts of compute.

Sherif Kamel: Saudi Arabia too.

Patrick FitzPatrick: Saudi Arabia’s building, Saudi’s building towards six. If you look at Egypt, we have one of the best geographies for the world when it comes to solar. And the mania right now is access to electricity to, to drive compute. And that's why data centers are rated in megawatts or gigawatts of power used. We have a ridiculous amount of the world's connectivity passing through Egypt thanks to the Suez Canal. Why should physical infrastructure not be part of the picture for us? Why are we talking about a few hundred megawatts of data center capacity being developed here, and not gigawatts being developed here?

Sherif Kamel: I believe the strategy that was developed in Egypt in 2021, and I think it was revisited in 2025, is looking at multiple dimensions, including ICT infrastructure, including policies and laws, including data, including talent. And if you go back to my model — which is people, data, governance, infrastructure — that bodes well with it. But of course, it needs a lot of investments, and it needs very specific targets. What are you going to do with it? And the word that keeps popping up all the time is that everybody wants to become "the hub of AI." Okay. If you put all that aside, these two countries have — and I mean by that, the UAE and Saudi Arabia — they use different approaches. So while the UAE had a head start, they focused more on deploying a governance model and more sophisticated model. If I look at Saudi Arabia, they are actually outpacing in terms of infrastructure buildup. Which one in two or three years will be more or ahead of the others? We don't know. AI is not a software. AI is an environment. It's a new team member in the football team. How can we use it? Is he or she better on the left side of the pitch or on the right side of the pitch? And should we use it throughout, or better to go back in the second half? I mean, I'm just making an analogy here.

Patrick FitzPatrick: Yeah!

Sherif Kamel: You need to understand how to be effective and add value at the age of AI, I don't mean by that the digital skills, but the complex problem-solving, the ambition, the curiosity, the forward-thinking, the communication as it should be are key components for a successful use and adoption of AI. So my question is: how much of those are effectively embedded in our educational curriculum?

Patrick FitzPatrick: Almost zero.

Sherif Kamel: Absolutely. That's the point. Are we teaching — and forget about AI for a second 

Patrick FitzPatrick: Sure.

Sherif Kamel: — are we teaching our graduates what they should know not just for the future of work, but for the work of today? Or are we still focusing on the standard curriculum?

Patrick FitzPatrick: We haven't been training people for the skills we need today in 30 years.

Sherif Kamel: 100%.

Patrick FitzPatrick: What does that say about the policy?

Sherif Kamel: It needs to be revamped. It's the content, i.e., the curriculum; it's how you deliver the curriculum; it's who is delivering the curriculum, and how you’re delivering the curriculum; it's how to think and not what to think. How to push students to be creative.

Patrick FitzPatrick: When you say revamp, what do we need to do to fix the policy? First off, what has to change in the strategy to make it workable?

Sherif Kamel: So let me go back to — and I go back in years to when the internet was introduced. I think I was fortunate and privileged to be working at a place where—I like to call it a as school — it was a magnet of young, dedicated, fresh graduates led by sort of a visionary team to how to not really plug and play the internet, but look at the local context and how the can internet be of use for Egypt, for Egyptians, for our society.

Patrick FitzPatrick: There was an active policy process where you were questioning that.

Sherif Kamel: Exactly.

Patrick FitzPatrick: Yeah. 

Sherif Kamel: So I claim that what works in Japan might not necessarily work in Egypt, vice versa. So the elements of the values, the culture, the mindset, the local context have always been on my mind. Who can introduce this culture/value/local context? It's not the machines, it's the individuals.

Patrick FitzPatrick: Absolutely.

Sherif Kamel: Fast-forward 30 years: the best AI strategy will be asking, "What do we want out of it?" 

Patrick FitzPatrick: Okay. 

Sherif Kamel: We can talk a lot of stuff, but it doesn't have to be a laundry list. We need to prioritize where to start. Having a strategy is important, but that strategy has to be through a number of phases where we learn as we do, and we adjust the following steps. So for me, education will come first, healthcare will come, agriculture will come—in that order. But if we go all-in at the same time, it's too much to absorb. We're still learning. The technology is still evolving. I see AI in the next 10 to 15 years contributing to GDP. And here, let me take a tangent with all those who are saying that technology is going to replace people. 

Patrick FitzPatrick: Yea.

Sherif Kamel: I mean, the World Economic Forum report, I think, last year said that 170 million, I think, jobs will be created and some of them will be replaced, and the like. I don't know about these numbers, but will jobs migrate? Yes.

Patrick FitzPatrick: Migrate where?

Sherif Kamel: Migrate to other jobs.

Patrick FitzPatrick: Okay, so the task will be done by somebody else.

Sherif Kamel: Or by a machine. 

Patrick FitzPatrick: Right. 

Sherif Kamel: But then those machines will require people to operate them. So am I saying that people are not going to lose jobs? No, but they're going to lose jobs to people who are better trained, better educated, more adaptive to the technology. So I'm going to be replaced by someone else, not by a machine. But other jobs will also be created. We are going through, again,  a shift in how we look at the work environment. So, I have a line that I keep on repeating: the world is all about change, and how you drive that change, and how you embrace that change. But the fact is, many people talk about change, push for change, promote change, deep inside, they do not change. 

Patrick FitzPatrick: They don’t want change, yeah.

Sherif Kamel: 100%. I think we should have that option. You're not changing just for the fact of changing, because there is a better way to do it, there is a more cost-efficient way to do it. And I think that's what technology in general has presented us over the last 40-plus years.

Patrick FitzPatrick: I'm grilling you about AI. I've had a number of other guests on the show talking about AI. Am I the outlier in thinking a lot right now about what AI means for society and for businesses? How much does this come up in your day-to-day conversations? Is it at the table when you meet with the Central Bank?

Sherif Kamel: It's everywhere. 

Patrick Fitzpatrick: It’s everywhere.

Sherif Kamel: There hasn't been a single meeting, meeting I mean board meeting or meeting to brainstorm, or a conference for the past maybe five years where the theme, or the title of the meeting, or agenda did not address AI and its implications — not only on business and education, 

Patrick FitzPatrick: Okay.

Sherif Kamel: But also on society, on employment, on migration, finding talent, on privacy, on security, on bias, on inequality. And I think that's a very important point, because more often than not, we tend to focus on the positives, which is great, but we also need to keep an eye on the negatives.

Patrick FitzPatrick: There are two debates there, right? 

Sherif Kamel: Of course!

Patrick FitzPatrick: There are people who say; on the one hand, all of this spending on AI goes to a small handful of a few dozen companies. On the other hand, we here in Egypt now have access to millions of hours of thinking and expertise that have been developed all around the world, which these AIs have been trained on. So what's the bear case and what's the bull case for us here in Egypt, specifically?

Sherif Kamel: For us, if we look at our needs, AI could be very useful for our society — the access to what technology can offer, but also access while being affordable.

Patrick FitzPatrick: So a lot of open-weight models.

Sherif Kamel: 100%. The beauty of it is — and I've seen that years ago with truck drivers who were probably illiterate — mobility was extremely important for them, and that was a very basic usage of the device. I mean, they used like 2% of the functions, and the device still made a difference to them. Imagine, especially with Arabic LLMs, what opportunities could be created. Yet there has to be a minimal level of awareness even before education, and the optimal scenario would be continuous, lifelong learning.

Patrick FitzPatrick: If we come back to the strategy for a minute, is the best use of AI tackling domestic problems first and foremost, rather than building industry?

Sherif Kamel: I believe in having an edge. So it's a societal, collective agreement that our edge should be in X sector, and then we put all our strengths there as a priority, and then we will address the other stuff. But where do we want our strength to be? You're saying manufacturing; someone else will say, "Why not tourism and hospitality?"

Patrick FitzPatrick: I think the thing that bothers me most about what you said, Sherif, is that it means government has to make some choices. We led a roundtable a few years ago where we had 20 CEOs in a room — high-profile members of our community — and we talked about FDI and exports and the need to choose a handful of priority industries. We had some interesting meetings with policy members about why it was necessary to choose, and the government's list was always 12, 13, 14, 18 sectors, because it's really hard to tell somebody, "You're not getting the export subsidy, no, I'm not going to give you a tax break." Not everybody can be a priority. How do we get government to make a decision and to limit the number of sectors it wants to prioritize today?

Sherif Kamel: That's a very difficult question, and I'll try to answer it the way I see it.

Patrick Fitzpatrick: Okay, yeah.

Sherif Kamel: Breaking the silos. The more you look vertically into silos, the more difficult you align at the top. A priority list cannot have 12 or 15 items. That's not a priority; that's a laundry list.

Patrick FitzPatrick: It's a laundry list.

Sherif Kamel: It's a laundry list. Priority at best would be like three.

Patrick FitzPatrick: Three, yeah.

Sherif Kamel: The less you have of silos, the more collective understanding you have at the top. It's understanding, it's structures, it's budgets, it's resources. You remove those barriers that separate them.

Patrick FitzPatrick: The barriers make it a zero-sum game. 

Sherif Kamel: 100%. 

Patrick FitzPatrick: Every dollar that comes to me, every pound that comes to me, every euro that comes to me is a dollar, a pound, or a euro that's not going to you.

Sherif Kamel: Exactly. If you remove that mindset — it's not an overnight fix, but that's the only way forward. AI thinks laterally, horizontally. 

Patrick FitzPatrick: Yeah.

Sherif Kamel: This is where technology actually could be of very good use. You look at the map that you have, at the potential, the prospects that you have, and how you can align all that in — again, I didn't say one direction, but three directions, three priorities. And they will feed into each other, they complement each other. And that's extremely important.

Patrick FitzPatrick: The reason today that Morocco has an aviation and avionics industry is because Boeing built a plant there to do wire harnesses for airplanes. And what the government did was map out in which industries there are very senior Moroccan expats around the world, and it led them to one guy at Boeing. They convinced this one guy to do the wire harness plant there. And now Airbus is there, Boeing is there, go across the list of industries in aviation and then all the adjacent industries in automobiles, etc. It starts with having people well-positioned globally who want to do something good for the country that they came out of.

Sherif Kamel: You look for your niche.

Patrick FitzPatrick: Yeah

Sherif Kamel: You look at what you're good at.

Patrick FitzPatrick: And Morocco made some hard decisions about that.

Sherif Kamel: Oh yeah. By the way, more often than not, the most successful decisions are the ones that are not, sort of, appreciated at the beginning because they're bold. You don't expect me to make the call where everybody is happy. But if that tough call makes you happy, that's the icing on the cake.

Patrick FitzPatrick: That's great

Sherif Kamel: But I'm not driven by just making people happy.

Patrick FitzPatrick: When we come back: the thing beneath it all... education.

———

Patrick FitzPatrick: You've said a bunch of times, Sherif, that your framework starts with people. I've challenged you on, "Are we graduating the right people with the skills the business community needs?" and you very rightly admit that we're not. What has to change to make it work?

Sherif Kamel: There is always room for continuous improvement. If I look at students today — and that's what we do with our students — we want to give them the best experience in the classroom in the discipline that they're having, but also outside the classroom, engaging in community development activities, which is completely different than what is being taught in the classroom. And then we have to factor in the experiential learning element that does not necessarily match the content of the discipline that is in the classroom. All this for me is exposure to the students, because at the end of the day, let's face it, they learn the bulk of what they do at work, on the job.

Patrick FitzPatrick: On the job. Yeah.

Sherif Kamel: But you need someone who understands the ecosystem beyond the degree...

Patrick FitzPatrick: You’ve got to teach them how to think, and you've got to expose them.

Sherif Kamel: ...and what is happening in the market, local and global. What are the dynamics? This is your soft power across different domains. These are your ambassadors in every walk of life.

Patrick FitzPatrick: You and I were talking last week, and you know that my daughter is just back from her first year abroad. And as she told you, she faced a faculty whose attitude towards AI is: if you use it, it is a disciplinary offense. Now, how is AUC approaching the question of AI and education specifically? And then within that, what are you doing at the business school?

Sherif Kamel: So there are two ways here: there’s a top-down and bottom-up. So the university itself at the top level has introduced a path, a direction where we need to capitalize on what AI can offer for the student learning experience. So, that's embedded in how courses are being taught. So that from one end is what the students get exposed to, but also from the other end, you need to get the faculty ready to teach those courses using AI. If I look at what we're doing at the business school, it's exactly the same. We’ve put together over a year ago, a task force that is represented by different disciplines to try to help our faculty and our students integrate AI in different shapes and forms into their learning experience. As part of our 2025 – 2028 strategic plan for the school, we have targets for where we should be by 2028. Do we have the adoption of AI across all the courses and sections today? No. Where we want to be by 2028 is a minimum of 50%.

Patrick FitzPatrick: So I'm allowed to use Gemini in the classroom.

Sherif Kamel: Oh yeah, absolutely, and it's encouraged. Will there be violations? Of course. Does it happen? Yes. Everybody's trying to outsmart the other, but I think it's part of the evolution of this new kid in town. And allow me here to just raise a question. So, I just came from one of the annual meetings for business school deans. We had a fantastic keynote speaker on the last day, and I told him, "Listen, in five years' time, how will you differentiate between a good scholar and someone who's very good at using AI?" And there was no answer.

Patrick FitzPatrick: Mhmm

Sherif Kamel: I think the answer will be how we orient and train and educate our learners on how to be responsible users of AI.

Patrick FitzPatrick: What's a hallmark of responsible use? How can you tell?

Sherif Kamel: To me, it always boils down to how you're being taught, the space you're given to express yourself and to showcase how you're using the different tools to reach the answer, and how you were oriented that it's okay to use it. Here we take us back to the three categories I mentioned at the beginning, where there is a group that says, "Wait and see." No, you can't wait and see. The train has already left the station and is going really turbo. AI is not going anywhere except it's going to get better, and I think the next generation of AI will be way more advanced than pushing a button. It may be moving and reacting based on how we think, how we react, how we move our hands.

Patrick FitzPatrick: That scares me.

Sherif Kamel: It scares me like hell! 

Patrick FitzPatrick: Yeah.

Patrick FitzPatrick: I've been talking with a lot of people in the service industry — lawyers, bankers, investment bankers, consultants, the people who employ your graduates. They are concerned about where the next generation of great lawyers comes from, because there's this universe where all of these folks train up as lawyers, bankers, and consultants by doing grunt work, by updating the model at 3:00 at night, by changing the damn slides, by taking notes during meetings, by grinding through contracts, or updating and applying changes to a contract. Where does that next generation of great bankers, lawyers, and consultants get trained if the AI eats up the work that's done by first and second-year associates?

Sherif Kamel: Yes, it's going to eat up some of the early automated, or possible to be automated, transactions, and that will leave room for more interpersonal communication and development. The things that a computer will do definitely will be automated/digitized. I think, like in banking, like in accounting, like in academia, what is the value proposition of professors today? To get into the classroom and give a talk for two hours? Absolutely not. The content he or she is going to share is already accessible to the students. But then if I go into the classroom, share a concept, discuss a case, and then we have a conversation for an hour or an hour and a half and let the students share how they see solving that issue, that's where the value proposition is. The one-to-many platform is no more. And I'd even take it to the next level: the notion of exams, whether essays or multiple choice or a mix between the two, in the age of AI, I don't think it's even matching.

Patrick FitzPatrick: Then what tests knowledge? What allows me to say, "I know what I'm doing"?

Sherif Kamel: Going back to the critical thinking part.

Patrick FitzPatrick: Okay

Sherif Kamel: We have a case, we listen to different opinions, and then we ask the students, "If you are in charge, what would you do?" I think that's the way forward. It's a co-open and learning space where you spend time in the classroom, where you spend time on campus, and the time you spend off campus.. It's all part of your learning experience. Are the faculty ready for that? Absolutely not. That's a transition as well. The faster we get into that new mode, the faster the students will learn the way they should be learning for the employers to be satisfied when they hire them. We used to call it "bridging the gap between business and industry" and so on and so forth. There is no "bridging the gap," it has to be seamlessly integrated. And I'll give you an example: at the school, we've started a business 360 course. 16 lectures, one faculty member, but 16 experts, people from industry and business. In 16 weeks, you're discussing issues in the same sector we’ve used, but from different perspectives. So you get exposed to 16 different experts.

Patrick FitzPatrick: Interesting.

Sherif Kamel: The exam is  the issues we've discussed at length throughout the sessions, what did you get out of it? If you were in that position, what would you have done? What decisions would you have taken? This is what I need to hire as an employer. 

Patrick FitzPatrick: Yeah.

Sherif Kamel: We have to understand that academia in general around the world is slow.

Patrick FitzPatrick: Part of that is probably by design and probably healthy for society at a certain level. AI transformation is hard to manage in any institution. What did you guys get right that listeners can take and apply to their business right now as they look at, "How the hell do I lead AI transformation in the business that I run or the department that I run?"

Sherif Kamel: I think the lesson we learned is, don't rush things. Don't be obsessed with coming up with a statement that "we're doing this and that." Just have a thorough discussion with different stakeholders, including students, by the way. And I always keep saying: those students, we enjoy having them for four years, or two years if they're graduates, but then they remain our ambassadors for 50 years. 

Patrick FitzPatrick: Yeah.

Sherif Kamel: We want to get out the best message and the best example possible so that we continue to make not just a different society, but a sustainable difference. So, if we rush it, it might not be sustainable. By the way, do we make mistakes? Of course. Do we adjust all the time? Absolutely. Did we get it right all the time? No, of course not. So we try to really focus on: what can we add not just for the sake of adding something new, but adding something that makes a difference?

Patrick FitzPatrick: As somebody who has made a career out of leading IT transformation, technological transformation, what is the biggest issue that you see a business leader facing in leading this AI transformation?

Sherif Kamel: What to prioritize and be driven by impact. What am I using it for? And that should not reside at the top level only. Aligning the different players in the conversation, even if they're not directly involved, is extremely important. So alignment is extremely important; it breaks the silos, but also rallies the different stakeholders in one direction. Communication is extremely important, formal and informal, because some of the informal channels help move the organization together. By the way, we don't have a problem of accessibility or availability of information, but what information to use, and how to use it, what information is actually real, and how can you make sure it's the right information to use?

Patrick FitzPatrick: That becomes the question. Yeah. Having the ability to choose where to point the lens, having taste; Do you see the machine ever having that ability?

Sherif Kamel: I sincerely hope not. You never know.

Patrick FitzPatrick: Yeah.

Sherif Kamel: But I stick to my guns that the human elements of today are going to be there for the foreseeable future. And that is what completes what AI can do. Think of it as your digital twin. It's there.

Patrick FitzPatrick: It's not going anywhere.

Sherif Kamel: It's not going anywhere. On the contrary, it's going to get better. 

Patrick FitzPatrick: Yeah.

Sherif Kamel: Let’s assume that the intention is always to get better, but in the right way, the responsible way, the way that does not widen inequality. On the contrary, it closes down the divides.

Patrick FitzPatrick: What inequality does AI help us address?

Sherif Kamel: You will think that I'm a broken record. I believe education and lifelong learning and training and professional development and awareness and all the synonyms that I can use, is a human right.

Patrick FitzPatrick: I won't disagree.

Sherif Kamel: Everybody deserves to get quality education.

Patrick FitzPatrick: It's the great leveler.

Sherif Kamel: 100%. So we need to be open and ready and pushing very hard to give everybody a chance to contribute to their own community, because the collective outcome of that is the country going forward fast. If you put that turbo engine, aka AI, I think the sky's the limit.

Patrick FitzPatrick: Are we able to regulate AI as a nation?

Sherif Kamel: Absolutely. And there are laws in the making that also relate to security and privacy, and countries are learning from each other. I would imagine that it's better to be safe than sorry, so I would expect that in the early editions of those laws, there would be stricter frameworks and parameters. Technology evolves because of the good uses of it, but also the misuses of it over the years that actually put checks and balances. But also, we need to put parameters so that intellectual property rights and privacy and security issues are in place.

Patrick FitzPatrick: You said something the other day that sort of stayed with me: "The more AI does for us, the more lonely we're going to become." Tell me about that.

Sherif Kamel: This must have been 20 years ago. We did a study of over a thousand, I think, students. And the average was, for an average of four people sitting together in a 30-minute span, they look at each other, talk to each other without sending a message or answering a phone or making a call was 90 seconds.

Patrick FitzPatrick: The longest period?

Sherif Kamel: That's the longest. That's 20 years ago! That's 2005, 21 years ago. 

Patrick FitzPatrick: Opa

Sherif Kamel: Imagine today I think it’s going to be like instead of 90 seconds, probably 20 seconds. So, if you're going to use the phone while you're sitting with your friends or with your family, you're lonely. Your friend is the screen. You'll be looking at the screen, on average, 16 hours a day. That's madness.

Patrick FitzPatrick: That's craziness. Yeah. What is the antidote?

Sherif Kamel: Be in charge. And let me reveal a secret here: I don't have a phone. Am I surviving? Of course. I think I'm more connected than most people.

Patrick FitzPatrick: How many emails did you get last night?

Sherif Kamel: Every day, 650. I'm connected. You choose. Maybe that was my way to control how to use technology. So I think the answer to your question is: have limits, have control. Turn on "Do Not Disturb," do whatever you want. I wouldn't dictate anything to anybody, but just be in charge.

Patrick FitzPatrick: Anybody who's used the free version of ChatGPT knows that it exists for one purpose: that last question is always a suggestion for you to engage with it further, or a question to make you answer it and continue the conversation. 

Sherif Kamel: Yeah.

Patrick FitzPatrick: My last question for you today, kind sir, is simple: what question have I not asked you that you wanted me to ask you today?

Sherif Kamel: The question would be, "What's next?" But I don't have the answer.

Patrick FitzPatrick: I don't think any of us do.

Sherif Kamel: Nobody has the answer. I see it as an incredible opportunity for Egypt. I still think it's our soft, yet most important power. This is an opportunity for a completely different new social contract, giving access to everybody and trying to get the best out of them. It's a very dynamic environment. Given that AI is not software, it's an ecosystem, a platform. Let's talk less about access and talk more about adoption, because access could be so many things. 

Patrick FitzPatrick: Sure.

Sherif Kamel: Last but not least, going back home to my field, I think the future of education at large, if we get it right, it will move from a transactional phase to a transformational phase.

Patrick FitzPatrick: Unpack that for me.

Sherif Kamel: Transactional is you getting a course, you’re getting a degree, versus a transformational learning experience. It's not about the classroom, it's not about the exam, it's not about the discipline. It's about the experience that you get exposed to, not just at the university level, but at the high school level. Kids are getting way smarter. I want to end with an interesting example here, if I may. 

Patrick FitzPatrick: Please.

Sherif Kamel: My daughter is 35 now. She must have been seven, even six, and we bought a TV at home. It was one of those new LCDs, blah blah blah. So my wife told me, "Why don't you make it happen, see how it works?" And I tried. Didn't know how. Apparently my daughter, again, she was six or seven, heard me say that, and in like 20 seconds, the TV was up and running. I heard her tell my wife, "How come he works in IT?" … Kids are smart! We can't keep teaching them the same way.

Patrick FitzPatrick: The kids are going to be all right. 

Sherif Kamel: Yeah.

Patrick FitzPatrick: Sherif, thank you so much for joining us today.

Sherif Kamel: Thank you so much for having me.