Generative AI: Its Rise and Potential for Society
Malcolm Gladwell: Alright, welcome everybody. You guys excited? Hello, hello. Welcome to Smart Talks with IBM, a podcast from Pushkin Industries, iHeartRadio and IBM. I’m Malcolm Gladwell. This season, we’re continuing our conversations with New Creators— visionaries who are creatively applying technology in business to drive change, but with a focus on the transformative power of artificial intelligence and what it means to leverage AI as a game-changing multiplier for your business.
Today’s episode is a bit different than usual. I was recently joined onstage by Darío Gil for a conversation in front of a live audience at the iHeartMedia headquarters in Manhattan. Darío is the Senior Vice President and Director of IBM Research—one of the world’s largest and most influential corporate research labs. We discussed the rise of generative AI and what it means for business and society. He also explained how organizations that leverage AI to create value will dominate in the near future. Okay, let’s get to the conversation.
Malcolm Gladwell: Hello, everyone. Welcome. And I'm here with Dr. Darío Gil. And I wanted to say, before we get started—this is something I said backstage: that I feel very guilty today because, you're the arguably one of the most important figures in AI research in the world, and we have taken you away from your job for a morning. It's like if, you know, Oppenheimer's wife in 1944 said, “Let's go and have a little getaway in the Bahamas.” It's that kind of thing. You know, what do you say to your wife? “I can't. We have got to work on this thing I can't tell you about.” She's like, “Get me out of Los Alamos.” “No.” So I do feel guilty.
Um, we've set back research by about four hours. but I wanted to—you've been up with that, with IBM, for 20— Dario: Years. Twenty years this summer. Malcolm: So—and how old were you when you—not to give away your age, but you were how old when you started? Dario: I was 28. Malcolm: Okay. So I want to go back to your 28-year-old self. Now,
if I asked you about artificial intelligence, I asked 28-year-old Darío, “What does the future hold for AI? How quickly will this new technology transform our world?” et cetera, et cetera, what would 28-year-old Darío have said? Dario: Well, I think the first thing is that even though AI as a field has been with us for a long time—since the mid-1950s—at that time, “AI” was not a very polite word to say, meaning within the scientific community. People didn't use, sort of, that term. They would have said things like, you know, maybe, “I do things related to machine learning,” right? Or “statistical techniques, in terms of classifiers,” and so on. But AI had a mixed reputation, right? It had gone through different cycles of hype and, it's also had moments of a lot of negativity towards it because of lack of success. Um—and so I think that that will be the first thing. We'd probably say, like, AI is like— what
is that? Like, you know, respectable scientists are not working on AI defined as such. And that really changed over the last 15 years only, right? I would say with the advent of deep learning, over the last decade, is when that reentered again the lexicon of saying “AI,” and that that was a legitimate thing, to work on. So I would say that that's the first thing—I think we would have noticed a contrast 20 years ago. Yeah. Malcolm: So at what point in your 20-year tenure at IBM would you say you kind of snapped into the present kind of “wow” mode? Dario: I would say, in, maybe the late 2000s. When IBM was working on the Jeopardy! project, and just seeing the demonstrations of what could be done in question- answering; it— Malcolm: Literally, Jeopardy! is this crucial moment in the history of AI. Dario: You know, there had been a long and wonderful history, inside IBM on AI. So, for example, in terms of, like,
these grand challenges at the very beginning of the field’s founding, which is this famous Dartmouth conference that, actually, IBM sponsored, to create, there was an IBMer there called Nathaniel Rochester, and there were a few others who, right after that—they started thinking about demonstrations of this field. And then, for example, they created the first, game to play checkers and to demonstrate that you could do machine learning on that. Obviously, we saw later in the ’90s, like chess, that was a very famous example of that. That was Deep Blue. With Deep Blue, right? And, playing with Kasparov. And then—but I think the moment that was really—those other ones felt like, kind of like brute force, anticipating sort of like moves ahead. But this aspect of dealing with language and question-answering felt different. And I think for us internally and many others,
was when—a moment of saying like, wow, you know, what are the possibilities here? And then soon after that, connected to the sort of advancements in computing and with deep learning, the last decade has just been an all-out, you know, sort of like front of advancements, and that—and I just continue to be more and more impressed. And the last few years have been remarkable, too. Yeah. Malcolm: I'm going to ask you three quick conceptual questions before we dig into it. Just so I sort of get a—we all get a feel for the shape of AI. Question number one is, where are we in the evolution of this? So, the obvious que—we, we all suddenly are aware of it, we're talking about it. What—can you give us an analogy about where we are in the kind of likely evolution of this as a Technology? Dario: So I think we're on a significant inflection point. That, it feels like
the equivalent of the first browsers when they appeared, and people imagined the possibilities of the internet—or more, imagined experiencing the internet. The internet had been around, right, for quite a few decades. AI has been around for many decades. I think the moment we find ourselves in is that people can touch it, and they can— before, there were AI systems that were like behind the scenes, like your search results, or translation systems. But they didn't have the experience of like, this is what it feels like to interact with this thing. So, that's what I mean. I think maybe that analogy of the browser is
appropriate because it's—all of a sudden it's like, whoa, you know, there's this network of machines, and content can be distributed, and everybody can self-publish. And there was a moment that—we all remember that. And I think that that is what the world has experienced over the last nine months. So, and—but fundamentally, also what is
important is that this moment is where the ease of—the number of people that can build and use AI has skyrocketed. So over the last decade, technology firms that had large research teams could build AI that worked really well, honestly. But when you went down into, say, hey, can everybody use it? Can a data-science team in a bank, go and develop these applications? And it was like more complicated. Some could do it, but it was more—the barrier of entry was high. Now
it's very different because of foundation models and the implications that that has— Malcolm: With the moment where the technology is being— Dario: Democratized. Being democratized. Frankly, it works better, for classes of problems, like programming and other things. It’s really incredibly impressive what it can do. So the accuracy and the performance of it is much better. Yeah. And the ease of use and the number
of use cases we can pursue is much bigger. So that democratization is a big difference. Malcolm: You say, when you make an analogy to the first browsers—if you, if we—to do another one of these time-travel questions, back at the beginning of the first browsers, it's safe to say, many of the potential uses of the internet and such—we hadn't even begun, we couldn't even anticipate. Dario: Right. Right. Malcolm: Exactly. So we're at the point where the future direction is largely unpredictable. Dario: Yes. Yeah, I think that is right, because it's such a horizontal technology that— the intersection of the horizontal capability, which is about expanding our productivity on tasks that we wouldn't be able to do efficiently without it—it has to marry, the use cases that reflect the diversity of human experience and institutional diversity. So as more and more institutions said,
you know, I'm focused on agriculture, you know, to be able to improve seeds, in these kinds of environments, they'll find their own context in which—that—matters that the creators of AI did not anticipate at the beginning. So I think that that is, then—the fruit of surprises will be like, why, we didn't even think that it could be used for that. And also, clever people will create new business models associated with that. Like, it happened with the internet, of course, as well, and that will be its own source of transformation and change in its own right. So I think all of that is yet to unfold, right? What we're
seeing is this catalyst moment of technology that works well enough, and it can be democratized. Malcolm: Yeah. The next sort of conceptual question: you know, we could loosely understand or categorize innovations, in terms of their impact on the kind of, balance of power between haves and have-nots. Mm-hmm? Some innovations, you know, obviously, uh, favor those who already have a—make the rich richer. Some—the—some, it's a rising tide that lifts all boats, and some are biased in the other direction. They close the gap between. Is it possible
to say, to predict, which of those three categories AI might fall into? Dario: It's a great question. A first, observation I would make on your first two categories is that it will be—both likely be true that the use of AI will be highly democratized, meaning the number of people that have access to its power to make improvements in terms of efficiency and so on will be fairly universal, and that the ones who are able to create AI, may be quite concentrated. So if you look at it from the lens of who creates wealth and value over sustained periods of time—particularly, say, in a context like business—I think just being a user of AI technology is an insufficient strategy. And the reason for that is, like,
yes, you will get the immediate productivity boost of, like, just making API calls and, that will be a new baseline for everybody. But you're not accruing value in terms of representing your data inside the AI in a way that gives you a sustainable competitive advantage. So what I always try to tell people is, don't just be an AI user; be an AI value creator. And I think that that will have a lot of consequences in terms of the haves and have-nots, as an example, and that will apply both to institutions and regions and countries, etc. So I think it would be kind of a mistake, right, to just develop strategies that are just about usage. Malcolm: Yeah. But to come back to that question
for a moment, to give you a specific— suppose I'm a, I'm an industrial farmer in Iowa with 10 million in equipment, and blah, blah, blah. And I'm comparing it to a subsistence farmer, someone in the developing world, who's got a cell phone, right. Over the next five years, whose, whose well-being rises by a greater amount? Dario: Yeah, I think, it's a good question, but it might be hard to do a one-to-one sort of like attribution to just one variable in this case, which is AI. But again, provided that you have access to a phone, right, and some way to be able to be connected. I do think—so for example, in that context, we've developed, we've done work with NASA, as an example, to build geospatial models, using some of these new techniques. And I think, for example, our ability to do flood prediction—I'll tell you an advantage of why we'll be a democratization force in that context.
Before, to build a flood model based on satellite imagery was actually so onerous and so complicated and difficult that you would just target to very specific regions. And then, obviously, countries prioritize their own, right? But what we've demonstrated is actually you can extend that technique to have like global coverage around that. So in that context, I would say it's a force towards democratization—that everybody sort of would have access if you have some kind of connectivity. Malcolm: That Iowa farmer might have a flood model. The guy in the developing world definitely didn't, and now he's got a shot at getting one. Dario: Yeah, but now he has a shot at getting one.
So there's aspects of it that—so long as we provide connectivity and access to it—that there can be democratization forces. But I'll give you another example that, that can be quite concerning, which is language, right? So there's so much language, in English. And there is sort of like this reinforcement loop that happens, that the more you concentrate—because it has obvious benefits for global communication and standardization—the more you can enrich like base AI models based on that capability. If you have very resource-scarce languages, you tend to develop less powerful AI with those languages, and so on. So one has to actually worry and, and focus on the ability to actually represent, in that case, language is a piece of culture also in the AI such that everybody can benefit from it too. So there's a lot of considerations in terms of equity about the data, the data sets that we accrue, and what problems are we trying to solve. I mean, you mentioned
agriculture or healthcare and so on. If we only solve problems that are related to marketing, as an example, that would be a less rich world in terms of opportunity than if we incorporate many, many other broader sets of problems. Malcolm: Yeah. Who do you think—what do you think are the biggest impediments to the adoption of, of AI as you would like—as you think AI ought to be adopted? I mean, if you would look, what are the sticking points that you would— Dario: Look, in the end, I'm going to give a nontechnological answer. The first one has to do with workflow, right? So even if the technology is very capable, the organizational change inside a company, to incorporate into the natural workflow of people on how we work, is—it's a lesson we have learned over the last decade is hugely important. Mm-hmm? So there's a lot of design considerations. There's a lot of, how do people want to work, right?
How do they work today? And what is the natural entry point for AI? So that's like number one. And then the second one is, you know—for the broad, uh, value-creation aspect of it—is the understanding inside the companies of how you have to curate and create data, to combine it with external data such that you can have powerful AI models that actually fit your needs. And that aspect of what it takes to actually create and curate the data for this modern AI—um, it's still a work in progress, right? I think part of the problem that happens very often when I talk to institutions is that they say, “AI, yeah, yeah, yeah, I'm doing it, I've been doing it for a long time.” And the reality is that that answer can sometimes be a little bit of a cop-out, right? I know you were doing machine learning. You were
doing some of these things, but actually the latest version of AI, or what's happened with foundation models—not only is it very new, it's very hard to do. And honestly, if you haven't been, assembling very large teams and spending hundreds of millions of dollars of compute—in sum, you're probably not doing it right. You're doing something else that is in the broad category. And I think the lessons about what it means to make this transition to this new wave is still in early phases of understanding. Malcolm: So what would you say? I want to give you a couple of examples of people in real-world positions of responsibility. Imagine I'm sitting right here. So imagine that I am the President of a small liberal arts college. And I come to you and I say, Darío, I keep hearing about AI. My college
has— I'm making this much money. If—that every year, my enrollment's declining, I feel like this maybe is an opportunity. What is the opportunity for me? What would you say? Dario: So it's probably in a couple of segments around that, right? All one has to do is, well, what is the implications of this technology inside the institution itself, inside of the college, and how we operate? And, can we improve, for example, efficiency? Like if you're having very low levels of, of sort of margin to be able to reinvest, is, you know, you run IT, you run, infrastructure, you run many things inside the college. What are the opportunities to increase the productivity or automate and drive savings such that you can reinvest that money into the mission of education, right?—as an example. Malcolm: So number one is operational efficiency. Dario: Operational efficiency, is a big one. I think the second one is: within the context of the college, there's implications for the educational mission in its own right. How
will—how does a curriculum need to evolve, or not? What are acceptable use policies for some of these AI? I don't think—we've all read a lot about like what can happen in terms of exams and, and so on, and cheating and not cheating, or what—are they actually positive elements of it in terms of how curriculum should be developed? And professions? Sustain around that. And then there's another, third, dimension which is the outward-oriented element of it, which is like prospective students, right? So, which is, frankly speaking, a big use case that is happening right now, which in the broader industry is called “customer care” or “client care” or “citizen care.” So—and this question will be— education. Like, you know, “Hey, are you reaching the right students?” Around that—that may apply to the college. How can you create for them, for example,
an environment to interact with the college, and answering questions? That could be a chatbot, or something like that, to learn about it. And personalization. So I would say there's, like, at least three lenses with which I would give advice, right? The— Malcolm: The second, let's pause on the second one though, because it's really interesting. So I really can't assign an essay anymore, can I? Dario: Can I assign an essay? Malcolm: Can I say, “Write me a research paper and come back to me in three weeks?” Can I do that anymore? Dario: I think you can. Malcolm: How do I do that?
Dario: I think you can. Look, this—so there's two questions around that. I think that if one goes and explains in the context, like, “What is it? Why are we here? Why are we in this class? What is the purpose of this?” And, one starts with assuming, like an element of, like, decency in people, or people are there, like, to learn, and so on, and you just give a disclaimer: “Look, I know that one option you have is, like, just, put the essay question and click ‘Go,’ and, like, and give an answer, you know? But that is not why we’re here, and that is not the intent of what we’re trying to do.” So first I would start with the—sort of like the norms of intent and decency, and appeal to those, as step number one. Then we all know that there will be a distribution of use cases—that people like that will come in one ear and come out of the other and do that. And,—so for a subset of that, I think the technology is going to evolve in such a way that, we will have more and more of the ability to discern—right?—you know when that has been AI generated, right? And, created. It won't be perfect, right? But there's some elements that you can—imagine inputting the essay, and you say, “Hey, this is like—it— .” And for example,
one way you can do that, just to give you an intuition, you could just have an essay, uh, that you write with pencil and paper at the beginning. You get a baseline of what your writing is like. And then later, when you, generate it, there'll be obvious differences around what kind of writing has been generated from the other. Malcolm: Yeah, but you've turned—it's—everything
you're describing makes sense, but it greatly—in this, in this respect, at least, it seems to greatly complicate the life of the teacher. Whereas the other two use cases seem to kind of clarify and simplify the role, right? Suddenly, reaching students, prospective students, sounds like I can do that much more kind of efficiently. Yeah, I can bring down administration costs, but the teaching thing is tricky. Dario: Well, until we develop the new norms, right? I know it's an abused analogy, but calculators—we deal, we dealt with that too, right? And, it says, “Well—calculator. What is the purpose of math? How are we going to
do this?” and so on. And we have— Malcolm: Can I tell you my dad's calculator story? Dario: Yes, please. Malcolm: My father was a mathematician. Taught mathematics at the University of Waterloo in Canada. And in the ’70s, when people started to get pocket calculators, his students demanded that they be able to use them. And he said no, and he—they
took him to the administration and he lost. So he then changed. Completely threw out all of his old exams. Introduced new exams, where there was no calculation. It was all like, “deep think,” you know. Figure out the problem on a conceptual level and describe it to me. And they were all—students deeply unhappy that he had made their lives more complicated. Dario: But it's to your point. That's the point. To your—
Malcolm: Point. Right. The result was probably a better education. Right. He just removed the element that they could gain with their pocket calculators. I suppose it's a version of that. Dario: I think it's a version of that. And so I think they will develop the equivalent of what your father did. And I think people say, you know what, it's like—these kinds of things, everybody's doing it generically and none of it has any meaning because all you're doing is pressing buttons. And like the intent of this was something which was to teach you how to write or to think or something. There may be a variant of how we do all of this.
I mean, obviously some version of that that has happened is like, okay, we're all going to sit down and do it with pencil and paper and no computers in the classroom, but there'll be other variants of creativity that people will put forth to say, you know what? You know, that's a way to solve that problem too. Malcolm: But this is interesting, because—to stay on this analogy—we're really talking about a profound rethinking, just—using a college as an example. A real profound rethinking of the way—there's no part of this college that's unaffected by AI, (a). (B), in one case, I've made everyone's job easier; in one case I've made—I'm asking us to really rethink from the ground up what “teaching” means. In another case, I've automated systems that I didn't think of. I mean, it's like,
that's right. That's all—it's not all—that's a lot to ask someone who got a PhD in medieval language and literature, 40 years ago. Dario: Yeah, but you know, I'll tell you a positive sort of development that I'm seeing. The sciences around this, which is, you're seeing—as you see more and more examples of applying AI technology within the context of like historians too as an example, right? When you have archival and—you know, and you have all these books, and being able to actually help you as an assistant, right, around that. But not only with text now, but with diagrams, right? And, uh, I've seen it in anthropology too, right? And, uh, in archaeology, with examples of engravings and translations and things. That can happen.
So, as you see in diverse fields, people applying these techniques to advance on how to do physics or how to do chemistry. They inspire each other, right? And they say, how does it apply, to my area? So once, as that happens, it becomes less of a chore of like, my God, how do I have to deal with this? But actually, it's triggered by curiosity. It's triggered by—you know, there'll be like, you know, faculty that'll be like, you know what, you know, “Let me explore what this means for my area.” And they will adapt it to the local context—to the local, you know, uh, language, and the profession itself. So I see that as a positive vector. That is not all going to feel like homework, you know? It's not going to feel like, “Oh my God, this is so overwhelming,” but rather to be very practical, to see what works. What have I seen others do that is inspiring? And what am I inspired to do? You know, what— what is—how is this going to help my career? I think that that's going to be an interesting question for, for, you know, those faculty members, for the students and professionals. Malcolm: Sorry, I'm gonna stick with this example
alone, because it's really interesting. I'm curious—following up on what you just said—that one of the most persistent critiques of academia, but also of many, of many corporate institutions, um, in recent years has been “siloing,” right? Different parts of the, of the organization going off on their own and not speaking to each other is a potent—is, is a real potential benefit to AI: the kind of breaking down—a simple tool for breaking down those kinds of barriers. Is that a very, is that an elegant way of sort of summing that up? Dario: I really think—and I was actually just having a conversation with a provost very much on this topic very recently, exactly on that, which is: all these, this appetite, right? To collaborate across disciplines. There's a lot of, attempts towards our goal, right? Creating interdisciplinary centers, creating dual-degree programs or dual-appointment programs. But actually, in—a lot of progress in academia, happens by methodology too. Right? Like
a new, when some methodology gets adopted—I mean, the most famous example of that is the scientific method, as an example of that—but when you have a methodology that gets adopted, it also provides a way to speak to your colleagues across different disciplines. And I think what's happening in AI is, is linked to that. That within the context of the scientific method, as an example, the methodology about which we, about which we do discovery—the role of data, the role of these neural networks, of how we actually find proximity of concepts to one another—it's actually fundamentally different than how we've traditionally applied it. So, as we see across more professions, people applying this methodology is also going to give some element of common language to each other, right? And in fact, in this very high-dimensional representation of information that is present in neural networks, we may find amazing adjacencies or connections of themes and topics in ways that the individual practitioners cannot describe, but yet will be latent in these large neural networks. We are going to suffer a little bit from
causality—from the problem of like, “Hey, what's the root cause of that?” Because I think one of the unsatisfying aspects that this methodology will provide is they may give you answers for which they don't give you good reasons for where the answers came from. And then there will be the traditional process of discovery, of saying, if that is the answer, what are the reasons? So we're gonna have to do this sort of hybrid, uh, way of understanding the world. But I do think that common layer of AI is a powerful new thing. Malcolm: Yeah. A couple of random questions that come to mind as you talk. In the, in the writer's strike that just ended in Hollywood, one of the sticking points was how the studios and writers would treat AI-generated content—would writers get credit if their material was somehow the source for AI? But more broadly, did the writers need protections against the use of—. I could go on. You know what? You probably were familiar with all of this. Had you been—I don't know whether you were, but had either side called you in for advice during that? The writers, had the writers called you and said “Dario, what should we do about AI? And how should we—that should be reflected in our contract negotiations?” What would you have told them? Dario: I—the way I think about that is that I divide it—I would divide it into two parts. Pieces. First is: what's technically possible, right? And anticipate scenarios, like, what can you do with voice cloning? For example, it is possible there's been, um, dubbing, right? Like—let's just take that topic, right? Around the world, there was all these, folks that would dub people in other languages.
Well, now you can do these incredible renderings; I mean, I don't know if you've seen them, where, you know, you match the lips—it's your original voice, but speaking any language that you want. That's the thing. So basically that has a set of implications around that. I mean, just to give an example. So I would say: create a taxonomy, that describes technical capabilities that we know of today and, uh, applications to the industry and to examples of like, “Hey, I could film you for five minutes and I could generate two hours of content of you and I don't have to, you know—then if you'll get paid by the hour, obviously I'm not paying you for the other thing.”
So I would say “technological capability,” and then map with their expertise consequences of how it changes the way they work, or the way they interact, or the way they negotiate, and so on. So that would be one element of it. And then the other one is like a non-technology-related matter, which is an element of—almost of distributive justice. It's like, who deserves what? Right? And who has the power to get what? And, and then that's a completely different discussion. That is to say, well, if this is the scenario of what's possible, you know, what do we want? And what are we able to get? And, I think that that's a different discussion, which is, as old as life. Malcolm: Which one do you do first? I think it is very helpful to have an understanding of what's possible and how it changes the landscape, uh, as part of a broader, uh, discussion—right?— and a broader negotiation. Because, you also have to see the opportunities,
because there will be a lot of ground to say, “If we can do it in this way, and we can all be that much more efficient in getting this piece worked on or this filming done....” But we have a reasonable agreement about how we—both sides—benefit from it, right? Then that's a win-win for everybody, right? So that's a—I think that would be a golden triangle, right? Malcolm: Here's my reading, and I would like you to correct me if I'm wrong. And I'm likely to be wrong. Uh, when I looked at that strike, I said, “If they're worried about
AI—the writers are worried about AI. That seems silly. It should be the studios who are worried about the economic impact of AI.” Doesn't, in the long run—AI puts the studios out of business long before it puts the writers out of business. I only need the studio because
the costs of production are as high as the sky and the costs of production are overwhelming. And—whereas if I don't, if I have a tool which brings, introduces massive technological efficiencies to the production of movies, then I don't need—why do I need a studio? Why would they be the scared ones? Dario: Or maybe—or maybe you need a, like, a different kind of studio. Or a different kind of studio. A different kind of studio. Malcolm: What do you mean? In this strike, the frightened ones were the writers and were the studios. Wasn't that backwards? Dario: I haven't thought about it. But the implications of it—it goes back to what we were talking about before. The implications, because they're so horizontal—it is right to think about it. Like, what does it do to the studios as well, right? But then, the reason why that happens is that it's the order of either negotiations or who first got concerned about it and did something about it—right?— which is in the context of the strike. Um, you know, I don't know what the
equivalent conversations are going on inside the studio and whether they have a war room saying what this is going to mean to us, right? But it doesn't get exercised through a strike, but maybe through a task force inside, the companies, about what are they going to do, right? Malcolm: Well—and to go back to your thing, you said the first thing you do is you make a list of what technological capabilities are, but don't technological capabilities change every—? I mean, they do. You're racing ahead so fast. So you can't—can you have a contract? Again, I'm sorry for getting into a little weeds here, but this is interesting. Can you have a—you can't have a five-year contract if the contract is based on an assessment of technological capabilities in 2023. Because by the time we get to 2028, 2028, it's totally different, right? Dario: Yeah, but like, where I was going is like—there are some, abstractions around that. It’s like, what can we do with my image,
right? Like, if I generally get the category, that my image can be reproduced, generated, contents, and so on, it’s like, let’s talk about the abstract notion about who has rights to that, or do we both get to benefit from that? If you get that straight, yes, the nature of how the image gets altered, created, or something—it will change underneath, but the concept will stay the same. And, uh, so I think what’s important is to get the categories right. Malcolm: Yeah. Yeah. If you had to—if you had to think about the biggest technological obstacle, revolutions of the postwar era—last 75 years—we can all come up with a list. Actually, it’s really fun to come up with a list. I was thinking about this when we were, you know—containerized shipping is my favorite. The green revolution. The internet. Where is AI in that list? Dario: So I would put it first. In that context that you put forth,
since World War II, undoubtedly, like, computing as a category is one of those trajectories that has reshaped, right, our world. And I think within computing, I would say, the role that semiconductors have had has been incredibly defining. I would say AI is the second example of that as a core architecture, uh, that is going to have an equivalent level of impact. And then the third leg I would put to that equation will be quantum. Quantum
information. And that’s sort of like—I like to summarize that the future of computing is bits, neurons, and qubits. And it’s that idea of high-precision computation—the world of neural networks and artificial intelligence and the world of quantum. And the combination of those things is going to be the defining force of the next 100 years in that category of computing. But it makes the list for sure.
Malcolm: If it’s that high up on the list, this is a total hypothetical. Would you—if you were starting over; if you were starting at IBM right now—would you say, “Oh, our AI act operations actually should be way bigger”? Like, how many thousands of people working for you? Dario: So within the research division, uh, it’s about like 3,500 scientists. Malcolm: So in a perfect world, would you, if it’s that big, isn’t that too small a group? Dario: Yeah. Well, that’s like in the research division. I mean, IBM overall, there’s tens of thousands of people working on that. Malcolm: We’re talking, we’re talking about—but I mean, like, so, starting from—first, so you have a—you’ve, we’ve got a technology that you’re ranking with compute and, you know, up there with it in terms of a world changer. Are we—so what I'm basically asking is,
are we underinvested in this future? Dario: No, but so—so yeah, it’s a, it’s a good question. So like what I would say is that I think we should segment. How many people do you need on the creation of the technology itself? And what is the right size of research and engineers and compute to do that? And how many people do you need in the sort of application of the technology to create better products, to deliver services and consulting, and then ultimately to diffuse it through, you know, sort of all spheres of society? And the numbers are very different, and that is not different than anywhere else. I mean, I mean, if you give examples of—since you were talking about, in the context of World War II, how many people does it take to create, an atomic weapon as an example. It’s a large number. I mean, it wasn’t just Los Alamos. There’s a lot of people in Oakland. It’s a large number,
but it wasn’t a million people, right? Um, so, so you could have highly concentrated teams of people that with enough resources can do extraordinary scientific and technological achievements. And that’s always—by definition, is going to be, uh, 1 percent compared to the total volume that it’s going to require to then deal with it. Malcolm: Yeah. But the application side is infinite, almost. Dario: That’s exactly—so that is where, like, in the end, the bottleneck really is. So, with, you know, thousands of scientists and engineers, you can create world-class AI. Right? And, you don’t need 10,000 to be able to create the large
language model and the generative model and so on. But you need thousands, and you need, you know, a very significant amount of compute and data. You need that. The rest is, “Okay, I, build software,” “I build databases,” or “I build a software product that allows you to do inventory management,” or “I build, a photo editor,” and so on. Now that product, incorporating the AI, modifying, expanding it, and so on—well, now you’re talking about the entire software industry. So now you're talking about millions of people, right, who are required to bring AI into their products. Then you go a step beyond the technology creators in terms of software and you say, well,
okay, now what? The skills to help organizations go and deploy it in the Department of, you know, the Interior, right? And then I said, okay, well, now you need like consultants and experts and people to work there to integrate into the workflow. So now you’re talking into the many tens of millions of people around that. So I see it as these concentric circles of it. But to some degree in many of these core technology areas, just saying like, well, I need a team of like a hundred thousand people to create, like, AI, or a, or a new transistor or a new quantum computer— It’s actually a diminishing return, right? In the end, like, too many people connecting with each other is very difficult. Malcolm: But, on the application side, it was just—think of our example of that college. Just the task of sitting down with a faculty and working with them to reimagine what they do with this, with this new set of tools in mind, with the understanding that the students coming in are probably going to know more about it than they do—that alone—I mean, that’s a, that is a Herculean people problem.
Dario: It’s a people problem. Yeah, that’s why I started in terms of the barriers of adoption of that. I mean, in the context of IBM, an, an example—that's why we have a consulting organization, IBM Consulting, that complements IBM Technology, and the IBM Consulting organization has over 150,000 employees. Because of this question, right? Because you have to sit down and you say, okay, what problem are you trying to solve? What is the methodology we're going to use? And here's the technology options that we have to be able to bring to the table. In the end, the adoption across, uh, our society will be limited by this part. The technology is going
to make it easier, more cost- effective to implement those, uh, solutions. But you first have to think about what you want to do, how you're going to do it, and how you're going to bring it into the life of this—in this context, a faculty member, or, uh, you know, the administrator and so on in this college, right? Malcolm: With that Hollywood, that, that notion, I thought, which was absolutely, I thought really interesting that, in a Hollywood strike, you have to have this conversation about—a distributive justice conversation about how do we—that's, it's a really hard conversation, right, to have in a—so this brings me to my next point, which is that you—we were talking backstage. You have two daughters, one in college, one about to go to college. Darío: That's right. Malcolm: So, they're both science minded.
Darío: Yeah. Malcolm: So tell me about the conversations you have with your daughters. You have a unique conversation with your daughters because your conversa—your advice to them is, is influenced by what you do for a living. Darío: Yes, it's true. Malcolm: Did you warn your daughters away from certain fields? Did you say, “Whatever you do, don't be”—you know? Dario: No, no, no, no. That's not my style. I mean, for me, no. I try not to be like, preachy about that. So for me it was just about showing by example of things I love,
right? And, things I care about. And then, bringing them to the lab and seeing things, and then the natural conversations of things I'm working on, or interesting people I meet. So, to the extent that they have chosen that—and obviously this has an influence on them—it has been through seeing it, perhaps through my eyes, right? And what you see me do, and that I like my profession. Right?
Malcolm: But one of your daughters, you said, is thinking that she wants to be a doctor. But being a doctor in a post-AI world is surely a very different proposition than being a doctor in a pre-AI world. Do you think—have you tried to prepare her for that difference? Have you explained to her what you think will happen to this profession she might enter? Dario: Yeah. I mean, not in like, you know, incredible amount of detail, but yes, at the level of understanding what is changing, like this lens of the—information lens with which you can look at the world and what is possible, uh, and what it can do, like what is our role and what is the role of the technology and how that shapes at that level of abstraction, for sure. But not at the level of like, don't be a radiologist, you know, because this is what we want for you. Malcolm: I was going to say, if you,
if you're unhappy with your current job, you could do a podcast called Parenting Tips with Darío, which is just, “an AI person gives you advice on what your kids should do based on exactly this.” Like, “Should I be a radiologist? Darío, tell me.” Like, it seems to be a really important question. Darío: Yeah. Malcolm:, Let me ask this question in a more—I'm joking, but in a more serious way, surely it would—if—I don't mean to use your daughter as an example, but let's imagine we're giving advice to somebody who wants to enter medicine. A really useful
conversation to have is, what are the skills that are—will be—most prized in that profession, yeah, fifteen years from now, and are they different from the skills that are prized now? How would you answer that question? Darío: Yeah, I think, for example—this is, goes back to how is the scientific method on, in this context, like the practice of medicine, going to change? I think we will see more changes in how we practice the scientific method and so on as a consequence of what is happening with the world of computing and information, how we represent information, how we represent knowledge, how we extract meaning from knowledge as a method, uh, than we have seen in the last 200 years. So therefore, what I would like strongly to encourage is not about, like, hey, use this tool for doing this or doing that, but in the curriculum itself, in understanding how we do problem solving in the age of like data and data representation and so on; that needs to be embedded in the curriculum of everybody. You know, that is, I would say actually quite horizontally, but certainly in the context of medicine and scientists and so on, for sure.
And to the extent that that gets ingrained, that will give us a lens that no matter what specialty they go with in medicine, they will say, actually, the way I want to be able to tackle improving the quality of care, the way to do that is—in addition to all the elements that we have practiced in our, in the field of medicine—is this new lens. And are we representing the data the right way? Do we have the right tools to be able to represent that knowledge? Am I incorporating that in my own—sort of with my own knowledge in a way that gives me better outcomes, right? Do I have the rigor of benchmarking to, and quality of, the results? So that is what needs to be incorporated. Malcolm: How, in a perfect world, if I asked you to, your team to rewrite the curriculum for American medical schools, how dramatic a revision is that? Are we tinkering with 10 percent of the curriculum or are we tinkering with 50 percent of it? Darío: I think there would be, a subset of classes that is about the method—the methodology. What has changed. Like, like, have this lens of it to understand. And then within each class,
that methodology will represent something that is embedded in it, right? So it will be substantive, but it doesn't mean replacing the specialization and the context and the knowledge of each domain. But I do think everybody should have sort of a basic knowledge of the horizontal, right? What is it? How does it work? What tools you have, what is the technology, and like, you know, what are the dos and don'ts around that. And in every area, you say—“That thing that you learn? This is how it applies to, uh, anatomy, and this is how it applies to radiology,” if you're studying that. “Or, this is how you apply, in the context of discovery—right?— of cell structure. And this is how we can use it.” Or “protein folding, and this is how it, it does—.” So that way, you'll see a connecting
tissue through, uh, throughout the whole thing. Malcolm: Yeah. I mean, I would add to that. It's also this incredible opportunity to do what doctors are supposed to do but don't have time to do now, which is, they're so consumed with figuring out what's wrong with you that they have little time to talk about the implications of the diagnosis. And what we really want are—if we can free them of some of the burden of what is actually quite a prosaic question of “What's wrong with you?” and leave the hard human thing of let me—should you be scared or hopeful? Should you—, what do you need to do? What—let me put this in the context of all the patients I've seen. That conversation, which is the most important one, is the one that's—seems to me. So like if I had to,
I would add, if we were reimagining the curriculum of med school, I'd like—with whatever—by the way, very little time. Maybe we have to add two more years to med school. But like a whole—that's not gonna be popular. But the whole thing about bringing back the human side of, yeah, you know, now if I can give you ten more minutes, how do you use that ten more minutes? Darío: But in that, in that reconceptualization that you just did is what we should be doing around that. Because I think the debate as to like, “Well, am I gonna need doctors or not?” is actually a not very useful debate. But rather this other question is “How is your time being spent? What problems
are you getting stuck?” I mean, I generalize this by like the obvious observation that if you look around in our professions, in our daily lives, we have not run out of problems to solve. So as—an example of that is, hey, if I'm spending all my time trying to do diagnosis, and I could do that ten times faster, and it allowed me actually to go and, um, you know, and take care of the patients and all The next steps and what we have to do about it. That's probably a trade-off that a lot of doctors would take—would take, right? Yeah. And then you say, well, to what degree does it allow me to do that? And I can do these other things and these other things that are critically important for my profession around that. So when you actually become less abstract, and like we get past the futile conversation of like, “Oh, there's no more jobs and AI's gonna take it, all of it,” which is kind of nonsense, is: you go back to say, in practice, in your context, right, for you, what does it mean? How do you work? What can you do differently around that? Actually, that's a much richer conversation. And very often we would find ourselves—that there's a portion of
the work we do that we say, “I would rather do less of that. This, this other part I, I like a lot. And if it is possible that technology could help us make that trade- off, I'll take it in a heartbeat.” Now, poorly implemented technology can also create another problem. You say, hey, this was supposed to solve things, but the way it's being implemented is not helping me, right? It's making my life much more miserable, or so on, or I've lost connection in how I used to work, etc. So that is why design is so important. That is why also workflow is so important in being able to solve these problems. But it begins by, you know, going from the intergalactic to the reality of it,
of that faculty member in the liberal arts college or a practitioner in medicine in a hospital and what it means for them, right? Malcolm: Mm-hmm. Yeah. What struck me, Darío, throughout our conversation is, um, how much of this revolution is nontechnical. ’Cause to say, “You guys are doing the technical thing here, but the real, the revolution is going to require a whole range of people doing things that have nothing to do with software, that have to do with working out new, new human arrangements”—talking about that, I mean, I keep coming back to the Hollywood strike thing: that you have to have a conversation about our values as creators of movies; how are we going to divide up the credit, and the— Dario: Exactly right! Malcolm: Like that’s a conversation about philosophy, and, Darío: it's in the grand tradition of why, a liberal education is so important in the, the broadest possible sense, right? There's no common conception of the good, right? That is always a contested, uh, dialogue that happens within our society. And technology is going to fit in that context too, right? So that's why personally, as a philosophy, I'm not a technological determiner.
Right? And I don't like when colleagues in my profession, right, start saying like, well, this is the way the technology is going to be, and by consequence, this is how society is going to be. I'm like, that's a highly contested goal, and if you want to enter into a realm of politics or the realm of other ones, go and stand up on a stool and discuss whether that's what society wants. You will find that it's a huge diversity of, of opinions and perspective, and that's what makes, you know, uh, you know, in a democracy, the richness of our society. And in the end, that is going to be the centerpiece of the conversation. What do we
want? You know, who gets what? And so on, and that is—actually, I don't think it's anything negative. That's as it should be. Because in the end, it's anchor of who we want as humans, as friends, family, citizens, and we have many overlapping sets of responsibilities, right? And as a technology creator, my only responsibility is not just as a scientist and a technology creator; I'm also a member of a family, I'm a citizen, and I have many other things that I care about. And I think that—that sometimes in the debate of the technological determinists, they start now butting into what is the realm of justice and society and philosophy and democracy. And that's where they get the most uncomfortable, because it's like—I'm just telling you like, you know, uh, what's possible. And
when there's pushback, it's like, yeah, but, but now we're talking about how we live. And how we work and how much I get paid or not paid. So that technology is important. Technology shapes our conversation. But we're gonna have the conversation with a different language. As it should be. And technologies need to get accustomed to—if they want to participate in that world with the broad consequences, hey, get accustomed to deal with the complexity of that world. Of politics, society, institutions, unions, all that stuff.
And, you know, you can't be, like, whiny about it. It's like, “They're not adopting my technology.” That's what it takes to bring technology into the world. Malcolm: Yeah, well said. Thank you, Darío, for this wonderful conversation. Thank you, to all of you for coming and listening. And, thank you. Darío: Thank you. Malcolm: Darío Gil transformed how I think about the future of AI. He explained to me how huge of a leap it was when we went from chess-playing models to language-learning models. And he talked about how we still
have a lot of room to grow. That’s why it’s important that we get things right. The future of AI is impossible to predict. But the technology has so much potential in every industry. Zooming into an academic or a medical setting showed just how close
we are to the widespread adoption of AI. Even Hollywood is being forced to figure this out. Institutions of all sorts will have to be at the forefront of integration in order to unlock the full power of AI thoughtfully and responsibly. Humans have the power and the responsibility to shape the tech for our world. I, for one, am excited to see how things play out. Smart Talks with IBM is produced by Matt Romano, Joey Fischground, David Zha, and Jacob Goldstein. We’re edited by Lidia Jean Kott.
Our engineers are Jason Gambrell, Sarah Bruguiere, and Ben Tolliday. Theme song by Gramoscope. Special thanks to Andy Kelly, Kathy Callaghan, and the EightBar and IBM teams, as well as the Pushkin marketing team. Smart Talks with IBM is a production of Pushkin Industries and Ruby Studio at iHeartMedia. To find more Pushkin podcasts, listen on the iHeartRadio app, Apple Podcasts, or wherever you listen to podcasts. I’m Malcolm Gladwell. This is a paid advertisement from IBM.
2024-05-30 11:00
