David De Cremer On AI Savvy Leadership | Episode #944 of Six Pixels of Separation

So who are you and what do you do?

I'm David Kramer. I'm the Dunton family dean at the Moore McKim School of Business at Northeastern University where I'm also a professor in management and technology. I run the business school where I'm trying to push the school into the direction of bringing in AI more in the business education so that we can train our leaders to be responsible but have an ability to use AI as a tool to create business value for all stakeholders.

That's great. You got a new book out called the AI Savvy Leader, which is a great read.

Thank you. It's indeed reflecting really what I wanted to communicate. How savvy do business leaders today need to be? Obviously they don't need to be computer scientists. So I don't want to ask my students, nor do I want to ask any business leader, hey, what are the three top large language models going to be in the next two years? That's not the purpose. It's really about how savvy do you need to be so you understand what AI can do for you and what it cannot do. And with that knowledge that you can make your business decisions, but in an augmented way, meaning you can uplift your workforce, you can create more value, and you can work faster and more efficient.

David, I'm a big nerd for this topic. I'm a big nerd for reading all of these books on leadership. And my takeaway when I finished the book was you could have just erased the word AI and called it the Savvy leader. Because I do feel like the components of what, what you're trying to bring forward, it just fits in a world where there's so much innovation and disruption. And I like the way you differentiate AI from other transformations. We'll talk about that. But it was more of a leadership how to be a modern leader than it is about AI.

I would argue, and I'm not disagreeing with you, the book is much more about leadership than it is about technology. As I just mentioned, the Sevenness, of course, has a focus on AI. So how well do you need to understand basically AI, but at the end of the day, and it's a very simple message, but it's unfortunately a message that's not really well understood today in the corporate world, according to me, and that is leadership is not going to change that much, especially business leadership. How you're going to run your company, how you're going to create value for all kind of stakeholders. That's not going to change the behaviors that were successful and efficient in the past. You're still going to be doing this if you want to be a good leader. But the only difference is you will have to use AI as a tool in these leadership practices. So just to give you an example, many people in business, and that's one of the reasons why I started thinking about writing this book, is there's so much in awe.

About AI.

Of course we're being bombarded with all kind of messages going from, oh, everything's going to change and it's not going to work out well for you to, oh my God, where are we going? Humans are going to be superhuman beings. There's so much uncertainty and at the same time so much certainty as well that everything's going to change and it's inevitable. So of course if you say these things in the business world, where it's about still mostly short term and profit, people will think like, oh, whoa, what's going to change here? Am I still relevant? And that's what business leaders have on their mind, am I still relevant? And the very easy conclusion to draw is I'm not a tech expert, I don't understand AI. I sideline, and this is what I see happening in companies everywhere across the world. Business leaders are not involved in the process of AI adoption and implementation. The result, unfortunately, is that most AI projects fail or they really don't provide the return on investments that the big tech companies are telling us because they all do it out of the goodness of their heart for society. And we're going to take benefit of all these values. But unfortunately it's not materializing. And from a technology point of view, actually it shouldn't even be a surprise because the value of technology reveals itself as a J curve. It's not just plug and play and immediate value. That's a lie in itself. But because people don't understand AI, that's what they think will happen. They have no expertise. So they sideline. And this is the big dilemma. That or the conundrum, basically you could say that leaders are being faced with. On one hand they have to adapt to AI and what's coming, and on the other hand they have to comprehend the phenomena already to make it work. And that is a difficult one. And that requires what I call an alignment between corporate purpose. What is your company for? Why do you have your organization? What is it that you want to achieve? Which are business questions and AI comprehension. Do you understand AI sufficiently so that you can see how can it bring that value to our corporate purpose? And that's the level that you need to achieve and then you delegate to the tech people. But unfortunately today AI enters after the budget is approved, they sideline and tech starts driving business.

Think there's a problem though in using this analogy? If we think about what AI was prior to the commercialization of OpenAI and seeing all these new GPTs come, seeing Google get much more aggressive, seeing Microsoft get much more aggressive, that perhaps what we were doing before in AI was very similar to what Web1 looked like in the sense of this excitement and this boom in this bus and this investment in infrastructure and then along comes a Shopify as an example and suddenly you don't have to spend hundreds of thousands of dollars for an online store. It's cloud based, you're not on iterative versions where you have different suppliers and now you're dealing with value added resellers and implementations and these big digital things. And that maybe as we record this midway may that suddenly this idea just goes away because of the scale and speed by which we have compute. The scale and speed of which we're seeing the iterative nature of. I don't, I mean I've been doing this a long time also I have never seen a product development like we're seeing. If we just said OpenAI. It's one thing I am not going to discount for the life of me things that Google, Apple and Microsoft and Salesforce are doing as well. And that was my feeling at the end. I thought I wonder if maybe we're wrong about this one because of now that it's probably eight months past when the book was written and published because the nature of the publishing industry we might be facing something very different. Have you thought about that?

Well, yes, I've thought a lot about it and I would say yes and no. So on one hand, look, we've seen so many bubbles to your earlier point and I mean if you just think about cryptocurrency blockchain people are not talking that much about it anymore.

That didn't change the day to day knowledge work.

No, but you see the excitement there. I mean I'm pretty sure that people reached out to you to talk about these topics because it's extremely relevant, it's going to change your life and then now it's not the case anymore. So I do think that we're creating a little bit of a bubble with AI as well in the sense that, and this is what I see in the business world, the fear of missing out and it develops into an AI's arms race. What we've seen. It's like, oh, I mean, it's inevitable, but it's going so fast. We have to use it. I don't understand it. I have no clue what it is about, but I have to do it. So the biggest risk today is not using AI. Now that paralyzes a lot of business leaders, but also a large part of our society because of our lack of understanding. But on the other hand, it's also because. Is it really true? I agree with you that this is a different animal. Because the way I always look at it is with the cars, with, for example, we had the first steam engine, then we had cars. We outsourced our body to say, so we don't have to carry our suitcases anymore. We can go everywhere. Physical labor becomes less important because we can outsource it now. Now we can outsource cognition. Oh, and that's a dangerous one because now we're at the verge of outsourcing body and mind. What's left? And that is existential. That's existential.

I wonder if it's as existential as this admission. And maybe where I'm looking at is this admission that what does AI do really well? I think what AI does really well right now, especially in the generative tools, is taking something that was at one point really delegated to somebody who had an advanced degree, who was able to excel in their career. However, they did it from standard to middle management to leader, and that suddenly their tasks day in and day out become somewhat repetitive. I'll take a personal example, David. I owned a marketing agency that was focused on digital. And people would come to us and say, oh, we need the web, we need social, we need whatever the world is. And I, in the industry, in my agency was seen as somewhat of an oracle. We have this information, these tools and this process to do it. But it was quite repetitive by its nature. Meaning if I'm doing it for you at the Northeastern University, or I'm doing it for Coca Cola or whomever, I could repeat that task. AI eats that for lunch. I mean, I see that in my day to day work. I see that at scale. And I'm wondering if it's more of this idea that it would be very hard for humanity as a workforce to see this moment in time where you realize, oh, all of this work I put in and all this work that I do actually can be outputted. It's not perfect yet, but it's, you know, it's, it's getting there for a lot of, a lot of These tasks, which I think makes it a bit different than we go faster and we can move from here to there, or we put wheels on luggage type of thing. Like it seems more scary to the individual. It's an affront.

Yeah, yeah.

It's a front to who they are, you know.

Of course, yeah, it is. I mean, but what I was saying is the combination now outsourcing the body wasn't that existential in terms of prep. It's the cognition in itself. But then you put it together. So if we outsource body and we outsource mind, what is left of a human? Yeah, that's, that's the issue. So we seem to be closing a loop here where yes, it makes sense when we hear about you're going to be replaced. Oh, you may have problem, you may have troubles if you don't upskill. Because yes, you can be erased. That's almost the way it sounds. This is in society, but again also in business. And to your earlier point, we forget almost to ask the question, what is it that AI is good at and what is it that it's not good at? Because we're being, we're in awe. Especially with ChatGPT now, with all the, the conversational large language models. It is amazing. But what people do forget is it's still AI. It's not general artificial general intelligence, although they think they're closer, but it's not true. It's just more sophisticated, elaborated, it works f faster, it's better, but it's still in the same region that we know. You still need to define the problem. So you provide the input the ChatGPT generates and you have content, but you need to edit it. This is how I always explain it. So in the middle is the AI, which is a tool. You define the problem, you interpret it. And this is for example, we see this in business education as well. It used to be write an essay, Mitch, write an essay, I'm going to grade it. And that's your course. Now I tell my faculty the essay is the input because they of the whole evaluation process at the university. So we've moved on because the generation itself you can't reward anymore. Of course you want people to have those skills because they use it as a tool. Now the thing is, instead of generating content transferred into knowledge, knowledge that's applicable knowledge that's relevant and useful and that's purpose driven to what you want to achieve. And that's going to be our assessment and education. So if you translate this to the business world, it's yes, AI is there, but it's a coworker now and it's a tool to use. So how am I going to implement this? So going back to my new book, it's really about. And to your earlier point again, yes, it's a leadership book. It's still about leaders. How can you create a culture or acquire a cultural change where teams get on board with what you're doing? Can you make AI meaningful to them with respect to the tasks they're doing? We replace the repetitive aspects, but that means as an organization, we're going to invest also in the jobs of the future and train you for it. But I also need to be visionary enough, so savvy enough to explain to you why I think it's meaningful and why it should be done. And I need to have a certain narrative so I can bring tech experts and non tech experts together. Because leadership is about establishing collaboration within your company. So if you can't have these two parties talking and no one else is going to do it besides you, you have a problem as a company. And that's what we're seeing. So that's that level of savviness. Coming back to our earlier point, yes. Do you understand what it is that AI can do great. And then what it doesn't do, and that's all very practical. And of course we also have a normative point of view where you could say, you know what, we created AI and what was its purpose? Let's go back to the earlier discussions in the 60s where you see there was, there's a purpose to AI. Oh, great. What do you mean there's a purpose to AI? Because that gets lost in all the heavy hitters, the tech companies that are, that are just throwing all kind of persuasive messages at us is it's serving humanity. So that also means you have a choice. Meaning, okay, if I don't have to use AI, you don't. So that means you have a choice. And sometimes if it's on par with a human, well, and this is what I call creating holistic value. We need to step away from just seeing AI as an efficiency promoter, as simply in terms of productivity. Yes, it can enhance productivity if implemented well, which again refers to the leadership behavior. But at the end of the day, it needs to be holistic value. Because humans are more than only efficiency. We're not just only performers. There's much more to it. And that's something that as a leader you need to understand because you need to drive a sustainable change. People need to keep using it and improve on it. For example, once you bring it into an organization AI, the model is not going to be perfect right away. You need feedback, you need communication, you need these cycles of feedback. You need to motivate your people there. It needs to be sustainable. So if you only say like hey, great, Mitch, the AI is here, it's going to change your job in this and this way and that's it. You're not going to keep working long on it. If you feel like you're being threatened completely in your job, your security, your salary, whatever, you're not going to help me and say, hey David, the model doesn't work. You're just going to blame the model. What happens? Inefficiency. Yeah.

It's a complex one because I think about the fact that for everything we say it can't do, I'm skeptical. I think it might be able to do for all the things we think it's a great copilot for. I think you're right. But it could be more. And this is where, where one of my psychological brains goes when I think about the structure of business. And the second one is to the idea of leaders in that normally when you're deploying a big transformational aspect, you're going from let's call it a web one platform to a more social web or you're adding an E commerce functionality or you're adding in, let's even be grandiose, you're going to add in some cryptocurrency. Whatever it might be, it only touches to a certain degree one thing. What I think leaders are challenged with here is that it will definitely impact most layers, most silos in the organizations or departments and most areas in one way, shape or form. So you could have it doing things like in the accounting department automating certain processes so that you don't have to get this audited by the bank or get back to the accountants. We have an AI that can now do this. You can have it replaced into a certain degree, maybe at the other replacement level, maybe junior copywriters and things like that. You can have it being more efficient as a co pilot at a mid level where maybe you don't need so many mid level people. But this is my, it's almost like this X, y weird access of it's impacting every level and every department. And this is why I feel leaders stumble the most because the idea of well, I'm just going to put in the marketing department, you could do that. But ultimately everyone else in every vertical seems to be jockeying for some type of mastery with these tools. That's to me, I think, where the complication and bigness of AI is.

I agree. So I just published a piece in the May, June issue of Harvard Business Review which was called Inclusive AI Adoption. Basically, if you want to create value across the board, everyone needs to be included. So to your point, I mean, this goes back also, how do you deal with data? I mean, we're very bad at data sharing and just making handshakes across the board in companies that the data can easily flow within the company. So I always compare it with an airplane that lands. It's full of data, they check the data. But compared to most other companies, it doesn't take three months that it goes from the ground up to the top to say, hey, there's a problem. No, no, this goes much faster. So in my book, I call this, you need to build flat communication cultures, because that's the reality. But flat communication cultures can only be achieved by inclusiveness and that everyone understands this. So, yes, you need to have a collaborative culture, which then leads to what I mentioned already, reducing silos as much as possible. And communication is key in this case. Again, as a business leader, you need to be savvy enough to be to develop actually kind of narrative that both or all departments understand.

Actually, is it possible that maybe leaders need to cool the jets in the sense of because it's moving at the pace it's moving and because we don't know what's happening. I remember I had a job before I started my agency in marketing. It was for a mobile content provider. And the CEO, who's a good friend of mine, said, we're marketing the unmarketable. Why? Because we're creating mobile content in a world at the time where there wasn't even interoperability with text messaging platforms. Meaning if I was on Verizon and you were on Sprint, I couldn't even text message you. And the deck of where the content was was a literal deck on your phone, which had no color, no mobile Browser. It was 1, 2, 3, and you could play little yes, no game or something like that. And I do wonder if we're asking leaders to take on this thing that is an inevitable, as we say, which is AI is coming. We knew mobile content was coming, mobile browsers were coming, smartphones were coming. But at that point, at this point, maybe it's not manageable because we actually don't know the scale by which it is. And let the tech people battle it out until we figure out what the use cases are.

I agree fully. And I mean cool the jets is I think not only for company leaders, but it's also for the tech industry. So I mean, and I don't mean in terms of, remember last year it was like, okay, the open letter where the companies six months take a pause. I mean six months is also an arbitrary number. I mean, why six months? But I mean, but it's my opinion there's no problem, there's no problem with AI knowledge generation and development. So we know more about what is possible and we develop quicker. With all that knowledge, I don't see any problem. It can go even faster. For me, the knowledge in itself, the issue is here, and that's when, that's what I like when you put it in the marketing context is when you start commercializing AI and purpose becomes profit, then we have a problem. Yeah, because why do we see the arms race that is happening now? So what you see on the market is there's also what I call so. So technologies appearing because people just want to sell, tech wants to sell and everyone knows like we're in a race. So I want to get it out as soon as possible. I don't care if it's just incremental or even useless, but it will be out there. And that was interesting that OpenAI recently then said, hey, I see all of these applications appear now, startups trying to sell it and based on ChatGPT. Well, they said don't waste your energy because more is coming and it's much more powerful and we're just going to take it away from you. So what happens? Tech startups throw even more so. So technology in the market because it's a race and it's about profit. That's what I said. Purpose becomes profit. That's one thing. The second thing is when cool down your jets for managers, especially for leaders is really. Yes, because again, and this goes back to the lack of comprehension of AI, what do you want to use it for? Because the only thing they see and feel is a sense of urgency. I mean the boards in the us they're the best example. Boards know nothing about AI, but they're there for governance and seeing, hey, is this the right kind of leadership for running our companies? But please tell me, what is this AI about? There's literally up to, I mean the majority of boards in corporate America are, are actually lacking that knowledge. So when you then tell them like, oh, you should actually look at the budget really, because that's your job. And See, hey, make sure you have money for the infrastructure, for adopting it, but also for training your people for this and this and this. And then it becomes like, hey, wait a minute, I thought this AI thing was going to do it. No, no, no, no, no, no. You need to bring it in the company. You need to make sure that the budget is spent wisely and that the leaders are doing their job. God, I have no clue. But I only know it's urgent. Yeah, okay. So you see, I think it's really in terms of cooling the jets. Yes. The question is really. Or the point that we want to make. Yes, we need some urgent education at that level for a business today. And from that point of view, because it's lacking, it's going too fast for them.

Yeah, I don't even know where we start with this. And I think that's a challenge for most leaders. I can look at this and say, you go from an iteration of ChatGPT to just image generation, to then video editing, to what we've seen recently with ChatGPT4O, the visual interface, where it's seeing and doing in a very impressive way. I mean, I'm not a Luddite. When I see these demos and what it's doing, these are leaps and bounds ahead of where they were a minor few months ago. Typically, we'd wait years for this type of new product release. And I'm curious again where it seems like the pressure of speed to do this when you can't define this. This is the insanity of the moment. We have to do AI. But when I speak to. I've spoken to AI experts and they can't even define AI cogently, meaning not cogently. I shouldn't say, that's not fair. They can't define AI uniformly. Where we all look at it and go, that's AI. I've had one AI expert basically tell me a definition that my response was, it just sounds like computers doing what computers do things faster and better than humans. My definition would be, you can tell it what you want it to do and it can do it on its own and explain to you back why it did it. That sounds like a more pure version of what AI is from. From my perspective, for me, it's. It's a real hard place to be. And I kept going back to that question as I was reading the AI savvy leader that what I still don't know that we can uniformly say, what is AI?

It's a good point. I mean, I define it as a computer that Makes decisions, Seems to be thinking in a way that seems intelligent. It seems.

But what about computers or calculators don't already do that? You know what I mean? That's the thing.

Yeah. No, no. But that's the thing. But a calculator is not a smart machine.

Well, it seems to be doing things that I couldn't do in a faster and more efficient way than I could.

Faster in efficiency. But then, you see, this is where. Discussion over the thinking now. And this is interesting because, you see, it's a very important point, because with the development of ChatGPT and now, for example, 4.0, now everything you see, people are literally talking about artificial general intelligence.

AGI. Correct. Yeah.

And that's an interesting one because that's also an issue of definition 100%. If you look at two, three years ago, artificial general intelligence was actually, wow, that's super intelligence beyond humans. That's like opening the door to singularity. Today, they don't define AGI like that anymore. Now AGI is. Oh, it's up to the level of human intelligence.

Yeah. It feels like the Turing Test, a little like AGI is the Turing Test versus before. I agree with you. It was this idea of we have something doing things that we can't comprehend why it's doing it. It's clearly more intelligent to us, to the point where I think it would have forced us to redefine what intelligence is.

Yeah, exactly. So, I mean, the Turing Test is a good example. Yeah, it basically. They softened it and they weaken it.

But to.

But why is this. Mitch? You know, this is. Well, marketing, commercial, reason behind it. There's a commercial. We're close now. Now we're close.

Almost there. Yeah.

How amazing we are. I mean, seriously, if you. Let's say, I would say five to seven years ago, you would be in China, where. I mean, in Shenzhen, where things are heating up so quickly. Okay. They would say artificial general intelligence. You know, they're a little bit more optimistic because they think everything's controllable. In that sense, they want to go fast. 10, 15, 20 years. When you go there now, or even last year, I was still in Singapore, living in Singapore. Then they would say, oh, five more years and we have general intelligence. So I was thinking, like, why. Because if you're really honest, again, chatgpt, like I said, the input and the output is still human. And it's amazing, and it's very impressive. I'm not downplaying it at all. But it's not that we suddenly entered a different era. Of AI compared to 10 years ago, what AI was, because we've just punched it. We're just. It's so much more sophisticated, but it's not a radical innovation from the AI that we knew. It's just made perfect almost at the level that we are. So in that sense, it's still as far away from artificial general intelligence today as it was 10 years ago, in my view. It's just.

Yeah, I'm thinking about that. I'm trying to think of.

So we're talking about the definition of 10 years ago.

Yeah, I'm thinking more like inputs. And so if we think about inputs and we go, what do we know? We know right now that it has hoovered in the entire Internet, including all audio and video, that they developed technology that was able to scrape YouTube and all of the audiobooks and have all of that information in it. So now we have a system that can operate, you know, with a lot of inputs, a lot of really credible inputs. The joke used to be, when I say used to be, a couple months ago, it was this idea that, oh, it's going to output pretty mediocre content because that's the input of what it had. Now I'm going, well, it's got everything in there now. So I'm thinking that the input is much better. Their compute seems to be a lot better. And so what's going to be the definition by which you and I would look at each other, David, and say, oh, wow, like this is something we haven't seen or we couldn't come up with. I believe it gets more clouded. Because what is it really doing? It's predicting. It does a letter or a word and it's predicting what the next logical one would be. But if the inputs are suddenly so staggeringly great and big and the ability for it to leverage that data comes in, I could easily look at you in all honesty and thinking like an intelligent human being and say, it could probably fool me, right?

Because it's read way more stuff than.

You and I could read. And you probably read more than 100% of the population.

I agree, it can, but that doesn't mean that it's generally important, right? So because this is the weakness of us as a human being as well. So let me tell you a few things. So I like what you're saying. I agree with it. If you look at the output, for example, it's so amazing that it has moved us forward, because I would say now because of what? Because like I said, it has access to everything. It's so fast. I mean, of course I would be stupid not using it because the content that it generates, in a way it puts us all above average, I always say.

Correct. Yeah. Very good.

So basically the average has changed. So we're the average of five years ago. Wow. We're all better now. If we use the tool in a way that it should be used, great. So it has moved us further. So we've improved. Great. Now there's, of course, a few problems with that as well, because if you're all above average, so you could raise a few discussions here, like, oh, does it then is it still worthwhile to get an education, Mitch? I mean, training your unique skills? I don't know. Because if we make it all equal, that also has economic consequences. Because now everything you create, because everyone can create, I'm not going to pay for that anymore because everyone can do it. So that's going to drop. I don't believe that it's going to. The prices are going to stay low because at one point people are going to want authenticity again, the real deal. So just to give you an idea, if. If you see a beautiful painting and you say, hey, I do want to pay $15,000 for that. And then I say, but, Mitch, I made it. I'm pretty sure you're going to say, well, I'm not going to pay more than 5,000. And that's.

I might not, but. And I think maybe because of how I see the world. So I think you're right that there might be that. But I do think that when it comes to that, what digital has done very well in the past decade, I think has validated this idea that you can have digital things that are scarce, which was a big leap forward from an economics perspective. And I don't think that that's going to diminish. Meaning fan fiction that was traditionally free is now stuff that people would pay a significant amount of money for. And you see this in the. In the actual market of people wanting to do that. So I get very. When it comes to comments like that, what I thought you were going to say is, Mitch, if I showed you the Mona Lisa and I said you could have it, it's just on a postcard or it's on a mouse pad, or it's here, okay. It's not the one that is there in France. I get it. There's a scarcity of the one of one that's fine. But as we output, what I'm seeing from a creative standpoint is the mediocre. Mediocre so our rate rises on that. But perhaps the real delta is from the people who are good, and the people who are good, who can take it from good to great are the ones that are going to be able. And again, that becomes a bit of a power thing that we already have.

Oh, well, yeah, I mean, that's another topic. And I would love to talk about that more. Like, is there democratization going on with AI? No way. Oh, no way. No. But everyone keep selling it like this. But again, coming back to your point, I see why you're saying this. Now, what is really interesting is that if you look at research, let's see what research says on this. So this is literally an experiment that happened that I was saying about a painting. People on average will devalue it. It's the same thing with skill. So the most recent studies show that if you show messages, and this is your Turing Test here, there's a message written and how empathy. Oh, empathic is this person who wrote the message. Then they said, oh, very empathic. And it turned out, oh, these. The most empathic messages were written by AI compared to humans. But then when the people were told, oh, AI wrote this, suddenly they evaluated it less empathic. So they. So our brain is so focused, of course, on it needs to be human, so that value is going to be there. So what I meant is, economically speaking, at one point, we're going to look back for that authenticity in a way you saw with Spotify. Remember, everyone can put music out there on YouTube or we put it out there, but then still only a few people. And really the authentic, the really good ones. Now I'm started to think, oh, are we going to see the same thing with Gen AI? But then I saw two things. Research shows that it's especially those with lesser skills that benefit the most from AI and Gen AI and not the ones who are so skilled. So if that would be true, wow, that's really flat across the line, which would be a pity and a little bit contrary to what you're saying. A second thing that I see is, and this is at the end of the day where AI is not there. It's not an active participant in our society. It doesn't understand implicit norms. I always give the example of this study. So I'm going to tell you this. There's a person who order, who makes a reservation at a steak restaurant. Person goes to the restaurant, is seated, gets the menu. Person leaves the restaurant and leaves a big tip for the waiter. My question, did the person eat a steak.

We don't know.

Yeah, but now you're thinking like an AI.

Yeah, yeah.

AI would say no because I didn't say it. You will say, yeah, likely, because it makes no sense. Why the person.

Right. Why would you leave a ticket?

So. So this is, this is going to be the big difference there. Where you want to look for the value, is it useful and meaningful? So when I said you generate content, it needs to be translated to knowledge. What is knowledge? In our society, we like to see it relevant, useful, meaningful. And those are still, at the moment, at the moment, human skills. And that's what I meant with has AI been uplifted to a next level of AI development? No, it's just been optimized, what we already know. Yeah.

I'm so curious to see what happens though, when leaders don't just think about deploying it for teams, but actually start using it at scale. One of the things I've found in the past couple months as I've used it, is it's changed how I think and how I operate and it's provided from a management level, you know, opportunities that I would have never imagined possible. I'll give you one quick example. I do a radio segment every week and typically we choose a news item and I will have my perspective because I'm the olds already. I have my own perspective for the work that I've done. Old guy who's got my own shouting at the clouds thing and I use it. I've created a chat bot in there that gives me the opposing perspective of someone like me. I basically ask you to ask, act like a media theorist like me, but give me the opposing perspective. And I have to tell you that it's given me the ability not just create better content, but it's given me so much more perspective and depth that it's made me a better. It's crazy. It's made me a better person. When I even have conversations like this, I'm able to see things that I didn't see before because I was stuck in my own echo chamber, my own ideas, my own thoughts. It's not perfect, but it's enough that I go, ooh. And how do I know the result? I know the result because when I do these interviews, I'll usually hear the disc jockey say things like, oh, or exactly, or I didn't think about that, or I'm able to counter their desire to want to counter me in a very intelligent way because I've studied it more broadly and then I apply it to business. I mean, it does apply to business. And I wonder, I do wonder when leaders start using it in this work, not to just replace the middle and the entry, but in their own thinking, how it might change. Because the truth is I do have very smart business leaders who have already been inputted in terms of their knowledge. And I can say it with things like if this is a business challenge, how would you approach us from the perspective of xyz? I mean, as someone who's learned the outputs, it's pretty impressive.

Yeah, it's pretty impressive, definitely.

So I think that that's also the other point is when you start getting it not to the people in the work that they're doing, but into the boardroom, into management, into actual thinking. It's pretty amazing.

Yeah, yeah, but I think, I mean, your example is, it's a great example, shows why being an AI optimist can also include being an optimist about humans.

Right.

Because what you just illustrated is the point that I made earlier. Augmentation. So you're saying you've got more perspective, you feel you're more connected maybe with the outside world. You basically have more perspectives to consider in a discussion. It uplifted you, it elevated you. So you used it as the, in the way that it should be used. But what is interesting here again, and this is to my earlier point, you use it, Mitch, because you see, you use it when you think it's relevant, when it's meaningful to the other parties who are listening or talking to you. And that's your human ability. So basically you could almost say that the 4o that level of sophistication of AI actually maybe 95% of the things we do AI can do, but it's that last 5%. Sure, that's what you did. So you feel like, yes, you get actually more knowledge. It's also the perfect teacher in that sense. It gives you a lot of knowledge, it gives you different perspectives. But the reason that you're saying I have different perspectives is because you know which problem to focus on or which discussion point to focus on and make it relevant. And that's the same thing as going from content to knowledge generation, you see, and that's the roles we're all going.

To be pushed in. Yeah, I guess I get cold in terms of how I feel about this when I say to myself, well, it's doing over 50 plus percent of the work and I can see a time in the not so distant future where it has that agent feature where it's on me and it's saying hey, today these are the three things you really. It's starting to take that other part where it's actually giving me, you know, it's providing inputs is what the easiest way I would talk about agents is. Instead of it being an output, it starts providing inputs to the human.

Yeah, exactly.

And my. When I say I get cold about it, I fear about it is because I. It does seem like it's coming a lot quicker than we're anticipating that component of it. So I hate talking about speed and technology because I find it so boring. Every time we start with technology in the 80s, everything's fast, fast. We know it's fast, so I don't. Putting speed aside, it's more of its ability that I think is so challenging. The example that I was thinking about for your book would be the fact that. So what do leaders do? The real core question of the AI savvy leader and my initial reaction before reading the book was they better do something because I would assume the smarter employees are already using it and they're using it to leapfrog team members. And they're probably using it in a very unethical way that probably breaches both of their employment and NDA agreements because they're probably inputting into GPT not realizing that they're putting IP into that engine. Can you talk a little bit about this race? Because this seems to be the big data race, the corporate one that I'm mostly interested in. What is it? What is an organization to do? You have these teams using it, you don't have one behind. We'll call it a firewall. A black box version of it is that the first step is getting this. Between getting this version inside your organization.

Yeah, yeah, yeah, I think so. You raised two important points that I quickly want to address. The first one was the input there. I just wanted to say something about it. I agree with what you were saying. It provides that input. But you get a little bit of cold. Me too. And you know why I get cold is let's say if it would do 95% of the things we're doing, we're going to start treating it almost as an oracle. So everything it says, that's what we're going to work on. And that's a problem because. Because by 2025, 10 to 15% of the data that we will work with worldwide has already been generated by AI itself.

Right.

Now that's a problem.

Synthetic data.

Yeah, I mean, I know you don't like to talk about fast. You just.

No, no, I see it, that's how.

Fast it's going and we have no clue because we can't catch up. I compare it to the mad cow disease in the 90s in the UK. I don't know. Remember that one? So how did that come about? Because a cow ate contaminated cow meat. So this is like AI is starting to use data generated by AI, but we don't know anymore whether it's right or not. So if we're going to treat it as an oracle, a big problem. But that's for us. Second point, what you just said. Yes, employees are using it. So there was a recent survey that showed in certain industries, especially financial industry, up to 50% of the CEOs don't want their people to use Gen AI. So what happens?

Yeah, they're using it.

Yeah, the company has no overview of it, but they're still using it, which leads to a bigger threat to privacy concerns, to regulation, because you have completely no knowledge anymore what's happening in your.

Company or what went where.

Yeah, yeah, yeah. I mean, and that's especially in the financial industry, of course, extremely important. The privacy there, the governance. So, and this goes back to these CEOs, the boards, what we've said, they're completely uneducated. They have no clue on why to use it. And here, and here's the, here's the advice that you're looking for. Should be, yes, you should, because you know what, AI is there to be implemented and can only get better, like I said, if you provide feedback. So you need to be able to experiment with. A company needs to have a work culture where there's enough trust, flat communication and collaboration so that you can test your AI. Because the perfect model doesn't exist right away. You have to work with it because models need to be modified and it needs to be modified also towards the purpose of your organization. So as leaders, what is important is to create what we call trust and psychological safety so that people know, yes, you can use it, but we can and we can experiment. So build that culture. Second thing is, as a leader, you need to be aware. And this is why I say ethical upskilling is more important than ever. Because at the end of the day, it's us who use it, the AI. The AI is not going to say, hey, I'm going to be such a badass now and I'm going to trick all of them. No, it's how we use it. Which means as a leader, you need to build a culture where you know, look, we need to follow this up as well, transparency is important. We need to discuss why we're using AI. So hence why the question is so important. Why are we using AI? As a leader, you need to be able to give an answer to this and then implement a culture where people know, hey, you have these teams to go to. If you're doubtful about can we use it, we can't use it. So ethics is a topics to be discussed. Governance is something that's openly discussed. You be need to create those. So these are the, are these new challenges? Yes and no. You had to do the same thing in the past already, but now it's even more pregnant than ever.

Yeah. How do you think they're doing? Like when you would look at this. And again, let's put aside the fact that, wow, it changes every day. And to even make that choice is a tough. I would not want to be a leader today having to even make the choice of where I'm going to place my bet on which engine, let alone having to deal with a workforce that you know could be operating and in a very offside way. You can have junior marketing people just doing basic copy that all you're really doing is inputting all your press releases and information without thinking twice about, I just had to edit it or rewrite something or come up with an idea. How do you, how do you grab this? It feels like we're asking people to grab a cloud of dust. Grab this cloud of dust and just go. It's like, good luck, you know?

Yeah, yeah. So it. Now it's not easy to be a leader today. To be honest, I don't think it's ever easy if you really want to be an efficient. But today, the thing is, and this is why I go back to the first point that you raised. Hey, David, this is a book about leadership. No, it's not about AI. And I said, yeah, because you know what? Leadership is not going to change suddenly towards the future. These behaviors are still going to be important, the same behaviors. But what is going to change is that can you be an effective leader now by means of these behaviors in creating value while using AI, that's how you're going to be evaluated now. And that's not only your own personal use of it. And you will have to use it because remember, lead by example. I see this so many times in companies that leaders, the top level implements technology but doesn't use it themselves. The impression that leaves with the rest of the company is like, I mean, they're not using it, why should I use it? But it's going to be to your point, how are you going to bring it into the company, how are you going to implement it and how are you going to create value across the board? Which means you have to work with these people. Again, it's the, it's building that culture and that is extremely difficult if you don't even understand yourself. Now, the biggest challenges are if you look at our executives, 45, 55 years old when they went to school, there was no way. People stare at me.

Look, when I present this stuff, David, I get really scary blank stars. More than I did in the early 2000s. When I talk about things like the Internet's going to be something, pay attention, you know?

Yeah, yeah.

These are really scared blank stares for sure.

Yeah, yeah. So. So in that sense, yeah. Again, coming back to your earlier point, remember, do we need a cooling down period? Well, yes, we do, especially because of educational reasons for our business leaders and also to be optimistic about it. Hey, it's not too late, guys. I mean, it requires a certain level of sevenness which I try to identify. That's the level that you need and then just be a good leader in combination with that knowledge. But what is also important, and this is a point that I wanted to raise, is the cooling down is not simply educational. There's also something we shouldn't forget is if we bring AI, we've already established AI is not going to bring immediate value. It takes some time. It takes experimentation, it needs feedback, communication. We need to optimize it and then create value across the board. Well, in a way, there's a difference between AI being developed in the lab as we see it that these big tech companies do, because there's no noise, there's no stakeholders. What's the biggest change when you bring it into a business? They're stakeholders, so the risk goes up. It's the same as, hey, I've got, look at the newest Tesla. It's amazing. It's amazing. And 20% of the time in the lab, it didn't function optimally. But I mean, we're so blown away by it. We all want it. It's our newest toy. Just imagine, Elon Musk now calls you just right now and says, hey, I'm listening to this guy here, Mitch, I just want to tell you, that tool that you want, that newest thing, it's on the road now. And I've got better news. Even 80% of the time, it's perfect in the traffic. Only 20% of the time, it kills people. Are we still going to then accept it? No, of course not. Because now we're talking about stakeholders here. So it also shouldn't go this quickly, it should be more gradually. And this is where OpenAI changed the business. Mitch. In my view, they are the first company ever to put AI on a market without from the lab directly into society. In 2022 when they put ChatGPT3 under pressures of the economic pressures because they wanted to be the first, they had to move on because they knew what Google was trying to do. They threw it on the market and saying hey, here it is, we just developed it, we just no one knows which data they used, how the model is use it and then two to three years later we're going to see now oh yeah, but you know, this happened, this happened, this happened. But will they be responsible? Of course not. So they made it. That's a big change in what's happening.

Now I want to talk a little bit about. There's a sub story to what you're talking about that I think is really important which is ensuring that these algorithms aren't being co opted. So we've talked a lot about the human factor, we've talked a little bit about how we input things, but there's still a core technology there and there's a lot of concerns especially in the polarization of the world, the political landscape and what's happening that originally I think it. There's a great book called Weapons of Math Destruction by Kathy o' Neill where she talked about prejudice and DEI and things like that. There's been an evolution of that to this point now. But this, if you have people who have very different value structures developing these algorithms too, that can be in and of itself another huge issue. I don't think we've seen a good display from government in terms of really wrapping their heads around as the regulation really seems to be more around copyright and who's the information is who and who should be compensated. Tell me a little bit about what do we do to make sure that the people who are building the core algorithm are doing it. We use the word alignment, which is how do we get it to be with human values? I don't love all of that because what is human values? Whose human values? Eastern, Western, spiritual, socioeconomical. We could go on and on. What do we do to ensure that the core has meaning to it? How do we do that?

I think it's such an important question and it's not being discussed enough because.

You'Re working a lot on this, right? You've got your center on AI Technology for Humankind. You even wrote a book called Leadership by Algorithm, which came out before the AI. So you're thinking about this now. It's here. If we have a forest fire, what do we do with that core stuff?

Yeah. So I always call it the bell between the reactive and the proactive parties. I mean, look, anything that has to do with the law, regulation is reactive because we put out a new law. After it happens, we're going to react.

Because something bad happens. Exactly.

And that's the government. Basically. That's the government. So if we're very rational, we couldn't even expect them to prevent it, to be honest. I know, and I don't want to sound like a pessimist here, but this is just.

We have to be real. I don't call it. I'm optimistic. We're going to be some. We're going to put some realism into the optimism. It's fine.

Yeah, yeah. So the proactive is of course, like, I mean, an open AI is proactive. If you listen to a Sam Altman two weeks ago, I don't care if I have to spend $50 billion a year. I'm going to get to that general intelligence as fast as possible. Again, the AI race. Now, going to the ethics that you said, let me take that statement again, I don't care. You know what, people were so enthusiastic about it. But to your point, the underlying question is, I don't care because it's a risk.

I'm curious, David, when he said that, my reaction, candidly, the experience, I have felt the most, if I could align it to another statement in the history of transformation, all this, it felt to me a little bit like when Bill Gates, I believe, said, we're going to have a computer on every desk and a people. Not. Not that this was more inflammatory by the tone of words, but the intent of it was. I didn't. I kind of took it like that. Like, sometimes we need braver statements.

Yeah, no, we need it. But I mean, but if you look at the track record of Sam Altman, you should. That's what I'm saying. If you dig a little bit deeper, you may also start thinking, because, let's face it, every company. And you saw this in the communication, especially like, I mean, Zuckerberg, any big tech communication company, and we're going to connect. We're going to connect people. And so with the cell phone, with smartphones, it started with this. And it's always for the good of society, but it was always very unclear what they Meant by the good for society. Well, I mean, I've never really heard Sam Altman explain what it is good for the society. It's usually the senses at the end and it's good for society. We're still suffering from the same disease, so allow me some cynicism here. So yes, I also interpret like you did, but I also saw it a little bit from the other point of view, which is probably my professional, My problem maybe of looking at it like this is okay, but what are you going to do with it? Because it's clearly, it's a competition, it's fast. Now to your point, how are you going to make sure that the people who work on it are going to use it in a way that doesn't kill us, but just helps us? And that starts already by just the first steps on working on it. It's not even there yet. So how do you envision this? And I know it's very naive, but of course I'm a business school dean. One thing is still education. So I wrote a whole piece on because of these challenges and we need to address them. Like you said, to bring the realistic point in, the optimism is yes, we will use it. So we also need to train ourselves in what I call ethical upskilling, simply the ethics in itself. Seeing moral dilemmas, just being aware, is there a dilemma? Yes or no? Can you discuss it? Do you have a narrative to discuss it? Do you have a jargon to identify it in the development process? These are simple skills, have nothing to do with technology, but. But simply with our own moral development. I mean, Oppenheimer was such a big movie, got an Oscar. So everyone was suddenly reminded again of like, hey, wait a minute, we had these scientists in the movie and they were maybe naive and they were thinking, hey, we're going to do something that we don't even know yet. Of course it's inspiring, it's amazing. And then when they found it, suddenly they realized, oh, but we have no control over it. It was just taken away from us.

A monster. It becomes the monster.

And it's in that time that Einstein said that with all these developments, what it comes down to will be the moral development of humans in how we use it. Well, I think it's the same time now to say that and to put it front stage and use it as a lens rather than as an afterthought. So that's one step. And for example, I'm trying to do this really in our business school. It's about educating responsible business leaders. That's why I used this example earlier from OpenAI when they said like 2022, we just throw it from the lap. Right. In society, you see, that's not responsible because you didn't use the lens of, can we do this? Actually, because look, there were cases, for example, where people, and this is happening today as we speak, probably, who are communicating with ChatGPT. Starting to be. It's almost like a therapist or a friend. Yeah. They're asking, hey, what is. How should I look at the world? Because I'm concerned about this. Give me more information. This is the oracle that we were talking about.

Well, this is my fear of the algorithm.

Yeah, yeah, yeah. So, I mean, people committed suicide, their cases known because they got so. So basically the ChatGPT started confirming their negative worldview. They got completely strangled with it and they couldn't escape from it anymore. So you see, these are things you need to think about. So if the past is a predictor of the future, that's why I was saying, when I heard Sam Altman say I don't care, I was thinking, yeah, well, you know what? I'm also looking at that dimension here.

So let's. Before we wrap up, let's talk a little bit about education, your role. You write this book. You can't be an AI savvy leader. If you're not going to have an AI savvy dean. If you're not going to have an AI savvy dean, who is using this tool or technology that we can admit. I have my own experience with education is forcing teachers and academics to rethink how we teach. I mean, if we had to change the world because of the calculator for math, this is a whole other level of that. What are you thinking? What. I mean, we have schools right now, currently, they don't allow the kids to use it. Forget. How do we do this in a world where it's getting really good? It's getting really good.

Well, I mean, the thing that I advocate is, like, I made very clear last week, for example, we started the AI day for faculty, which is providing resources, accessibility, because, look, a lot of people have to learn about it. So you need.

That's it. We're asking them to do the impossible. It's like, it's like Covid all over again where, like, you shut down, like, you use Zoom, you know, like, okay, well, it's not so simple.

Yeah, exactly. So, but my. So I made clear right away, students have to use it. They have to. So. Wow.

Well, they're going to.

They're going, yeah, but again, same point as in the finance industry, they want you to use it, they still use it, so they have to use it. Which already, as I already explained, that's going to change how you're going to assess them and how you're going to evaluate them. And because that's going to change, it's going to change how you're going to teach. So basically they're already going to get knowledge. Just like you have knowledge when you talk to people and you just ask chatgpt beforehand some questions but they're going to bring this in. So basically professor, in the classroom the role is going to be a facilitator and a discussant towards self development, towards that awareness. For example, that ethical upskilling fits there that you're going to say like okay, so for example, I just wrote a business case on profit versus purpose in the tech industry with the, with the arms raise that we were talking about and that it meant like okay, what kind of output and economics can we expect if you're profit driven only or purpose driven only? Okay, where do you find the balance? Let people think about those things. But you see, these are going to be the bigger questions which relates to the skill of critical thinking. So you see already the role has changed. It's not anymore. I walk in and I said I'm going to lecture for two hours and that's it. No, those days are gone. So I make that very clear. And this is the right.

So help me understand this for people who are cured. Do you have the right staff for this? Do you have the right student for this?

That is, that's a perfect question. That's the question. The question. We have a large generation basically in amongst faculty.

We can't expect them to.

Yeah, no, but this is why I'm also seeing myself as, I'm actually transformational leader at the moment as a business school dean. So I'm being invited everywhere by my business school dean colleagues now. Oh, you wrote this book. Can you modify the title and saying what does it mean to lead a business school in the AI era? Yeah, and it's similar. So when I said experiment, provide a safe environment, still reward human involvement, make sure you bring the techies and the non techies together. Well, that's what I'm literally doing in the business school. So just few quick examples. So all our stuff, I created a data day. So why do data matter? How do you create handshakes between data? What kind of questions do I want you to ask? Literally feeding it faculty okay, you have to work. AI is a tool for business. If you want to understand which large language model in the next five years, you go to computer science. So that's not our job, but it's a tool. So how are you going to bring it in the classroom? But we have to give examples, examples and repeat, repeat, repeat, which means you develop incentives as well to promote that use so that they start understanding. But again, to our earlier discussion, of course, you can't do it overnight. It's.

You have tenured professors.

Yeah, I mean, it's a slow process. And that's.

So, so help me with this question. How are you then denoting value to the, to the parents and the students to pay for this when you know that you have teachers that you know. My analogy is the car analogy. You're selling a Mercedes, but you perhaps might be having teachers that are kind of Toyotas from five years ago. They're great, they drive, they're fantastic. But in this new world, that's not what you're paying for.

Well, and that's a great question. As you know, I'm not American, so I'm half Belgian, half Dutch, so I'm European. I lived in the Asia a long time. So for us, the fees that you pay at university are staggering. We don't know this because I'm Canadian.

So I got you too.

Yeah, so. So I, So when I started here, it was very funny because American faculty look at me. But you're not American. I said, no, that's a good thing. Start to change here because, yeah, you need to become aware that they're paying a lot of money, these parents, because a lot of professors don't think like this. They have their own world. So you really need to bring this to the table to say so. Yeah, and I'm, I'm very sensitive to it because my parents were able to let me go to the university because they paid at that time, I think it was a hundred dollars a year. I mean, I don't even know. I mean, it was just. But. And my alma mater is ranked number seven in the U.S. by the way, when it comes down to technology degrees. So, hey, we did it well, so how can you make that salient that they have a responsibility there as well? So that's one thing that I'm trying really to put on the table. And maybe because of my cultural background, I'm much more savvy or sensitive to it. That's one thing. The other thing is it relates to your earlier comment. It's not easy to Be a business leader. Well, let me tell you, it's not easy to be a leader in higher education either.

We didn't even talk politics. We left that out.

So this is indeed a challenge. But I can tell you at Northeastern, I mean, it's amazing. Simply like I said, our president has been talking about this for years on digital literacy, technology literacy and human behavioral literacy to make it meaningful. Remember our earlier discussion? Yeah. You can generate all of this, but now you're going to go work with it. That's, that's humans. So we're developing with lightning speed, I could say, I mean, this happened within two months, an AI certificate across colleges. So for all the colleges that basically these are the same courses, then each college is also trying to identify courses. I mean, we have business analytics combined with data analytics. We bring in the AI ethics where because Northeastern works with what we call a co op system. So that means they go out and they work for six, work for six months and they get an actual salary. So the great thing about it is why Northeastern has gone up so quickly is you have a job almost if you graduate because they have one to one and a half year work experience, which is amazing in today's economy. Now we're trying to prepare them. Now this is another initiative where we try to give them an AI certificate already. Oh, they've done these two or three courses. So when they go for their first job, for their first co op, that they can show this to employers as well. Like, look, we're being prepared to use it as a tool. And that's why I'm saying for the business school, it's a tool for computer science. Of course it needs to be a little bit more elaborated, but for us it's a tool. What does this mean? And that's actually in line with what I'm trying to describe in my book that they know. Well, that's how I look at it and that's how I'm going to create value for you. So that's what we're trying to do and it has to go extremely fast.

So tell me the one thing that made you think differently?

Oh, the one thing. Well, I worked and lived on four continents, speak five languages. And what made me think differently is I didn't need AI to your point, to create different perspectives. Culture did that job and that opened my eyes.

I think it's beautiful in the day and age we're in, David, for you to have this role and for us to be looking at how culture is impacting our world, how it's impacting politics. I think the school is really lucky to have someone like you with that perspective. That's great. The new book is called the AI Savvy Leader. David, let people know where they can find out more about the book you're writing and all that other stuff.

Well, go to my personal website, www.daviddechramer. one word. You'll find everything about me, my writings, or order the book on Amazon or any preferred online provider.

Thanks so much for your time. David.

Thank you so much, Mitch, for having me.