Chicago Booth Review Podcast AI Can Do the Work. Can It Build the Trust?

Hal Weitzman: Companies are rushing to develop AI strategies, encouraging their employees to become more productive and more efficient. But what is being lost in the process? Welcome to the Chicago Booth Review Podcast, where we bring you groundbreaking insights in a clear and straightforward way. I'm Hal Weitzman. Today, we hear from John Burrows in the second of our podcast episodes about his new book on social capital. What are companies missing in their AI strategies and how can we focus on where humans can add genuine value? John Burrows, welcome back to the Chicago Booth Review Podcast.

John Burrows: Good to be here, Hal.

Hal Weitzman: This is the second of our conversations about your book, Social Capital at Work: Building the Hidden Asset that Drives Trust, Engagement, and Performance, which you've written with Seth Rachlin, which is a great read, and I heartily encourage people to go out there, buy it. I guess you care more about them buying it than reading it, but you want them to read it as well and act on it?

John Burrows: Yes.

Hal Weitzman: Yeah. Okay. Buy your own copy, maybe is the way to put it. You have a lot of stories in this book, which makes it so entertaining. And one of the stories that really grabbed me was this one about the oil importer and the map room. Tell us about that.

John Burrows: Yeah, so this story is not my story. So I co-authored the book with Seth Rachlin, and we've woven multiple stories into the book. But this one, the map room story, is his. So a little bit of background on Seth. Seth for many years was a management consultant. He ultimately ran Capgemini's insurance vertical, so quite a successful management consultant before retiring and moving to academia.

But early in his career had a really big blunder around this opportunity. So he was brought in by an oil shipping company, and was told, "Make our process more efficient. Use technology to make things better." The situation was that they were ships moving around the globe with oil loaded on them that were coming to various ports. And you had to ensure that the port at the time the ship arrived had the right people and machinery to offload the oil. Seems pretty simple.

How did they track this? Crazy. They had a whiteboard with a image of the world, a global map, and then magnets, heaven forbid, being moved around on this board to show the position of the ships. How do they get the position of the ships? Phone calls. So there was a regular phone call between each of the captains of the ships back to headquarters.

Seth sees an opportunity here, right? This is stupid. Let's get rid of the phone calls because there's a thing called GPS now. And we'll just put GPS transponders on each of the ships, and we'll know where the ships are and end of story. Everyone got excited about this. Got rid of the pin up of the whiteboard and the map, and replaced it with a big shiny monitor. It looked great.

Problem is a few weeks later, ships were turning up early or late. And there was a sense of what was going on, what was missed? What was missed is Seth had taken the information out of that exchange. He thought those calls were simply conveying the location of the ship, but actually what they were conveying was other ancillary details, like potentially the ship's crew are restless this week, or some of them are sick, or we're having a bit of a problem with the engine.

There was other information being conveyed beyond merely the location of the ship. And that information, that sort of context, was now completely lost. We automated a conversation with a GPS signal, but stripped out a lot of information that we at the time didn't realize was crucial, but was actually crucial. And I would argue that's what a lot of us are doing right now with AI implementations.

Hal Weitzman: Right. I mean, you get to that at the last chapter or so of the book, this idea of lost context. So what are we losing when we're trying to get to a system that is apparently going to make us more productive? But who knows? I'm going to ask you whether that will happen. So let's start with this. The book is called Social Capital, but right in the last chapter you talk about AI capital. So can we stomach another capital? Tell us what AI capital is.

John Burrows: Yeah. So AI capital is, you might think and say, well, AI is just a tool, but I think AI should rightfully be considered a capital because in some sense it behaves as economic capital does in many ways. We can invest in it. It depreciates. There are returns to it. I think it can also usefully substitute for human capital. Think of many of the things that we all do. We produce PowerPoints, we write memos, we create financial models of an acquisition target. These things are all historically things that smart humans did leveraging their human capital. And I think many of us now recognize that the frontier models are quite capable of producing very robust versions of those various artifacts and deliverables.

And I think what we want to think deeply about is to what extent AI is substituting for human capital, but what's left? And I'm obviously partial to the idea of social capital, that social capital is those connections around and between people that augment, and the human capital we each bring individually to the table. And so my sense is we need to grapple with the real possibility that human capital is being substituted for AI capital, which dramatically increases the returns to social capital, I would argue.

Hal Weitzman: Who's doing this substitution? Is it individuals or is it institutions?

John Burrows: It's a bit of both. But here's an example. I'm guilty as charged, so do as I say, not necessarily as I do, although I'm trying to get better on this front. I'll give you a problem I had a couple of weeks ago. I was struggling. I was facing a deadline, and I needed a quick edit of a syllabus. And I needed to come up with a new assignment for an audience I was teaching in Saudi Arabia.

And I could have asked one of my TAs to do this, but in reality, that conversation would've taken too long, and I probably would've had to get multiple drafts back because the chance of my TA getting it perfect with round one was small. So what did I do? I asked Claude. My friend Claude or my friend ChatGPT. I got an instant response that was pretty good, but I didn't have that conversation with my TA.

And that conversation with my TA wouldn't only have been about please do this for me, but it also would've been an opportunity to build trust and rapport with her. Give her an opportunity to say, "John, I'm kind of struggling with some things." Or, "Here's what I've heard from a couple of other TAs." I have a fairly large team of TAs because I teach a lot of classes. "I've kind of heard a couple of rumblings from a couple other TAs who are struggling with this or that." The point is there are these other things that that interaction was doing that we ignore to our peril, I think.

Hal Weitzman: In a corporate environment, that means what? That AI sort of reduces the amount of on the job learning that people are doing?

John Burrows: Yeah. I mean, think about this. I mean, I think that the services industry, whether that's consulting or law, is really going to have to solve this problem. Because you think about what is the model by which one becomes a good young lawyer at an elite firm or a good young management consultant. It's an apprenticeship model. We hire you because you have the raw capabilities we think to learn by doing. Talk to older, wiser people who know more about this, get feedback from them, and over time learn to do this better.

Now what happens to senior people in law firms and consulting firms and beyond is we often default, not because we're trying to do anything destructive, we default to just this once I'll ask ChatGPT for the quick answer. But on aggregate, many of us making individual decisions that aren't necessarily trying to subvert anything actually end up unintentionally robbing the juniors of these important interactions through which they learn and get better.

And I think there's a non-trivial risk that we're going to have a dearth of capable people in the not entry level roles, but in those sort of five years in, does five years in from now to five years count the same as the five years past? And I think the answer's no, unless we step in and do something about it.

Hal Weitzman: So if individuals are all quietly doing their own thing, and that aggregates up to something that's systemically negative, does that mean it's the responsibility of companies or organizations to change that? Because one thing that I do want to ask you about is you portray social capital as an asset. In other words, it's an asset both for individuals, but for organizations, it helps stuff happen.

John Burrows: Yes.

Hal Weitzman: So how should organizations account for that asset and make sure they don't lose it because of all these small individual decisions that you describe?

John Burrows: Yeah. I mean, look, I think social capital should be treated as an asset and thought of one and invested in, but the focus historically, and remains, on human capital and building human capital initiatives. And I'm not here to say we shouldn't continue to invest in our own human capital and the human capital of our employees. And I'm not suggesting getting an MBA or a PhD is a bad use of time. But we should recognize that your resume, human capital gets you the job, gets you in the door, but it doesn't get your work accepted.

And as I say, more and more to the extent I use AI to do these tasks, who's going to stand behind that memo? Who's going to deliver that PowerPoint? Who's going to say, "Hal, trust me, I'm going to shake your hand. I've got your back. If this goes wrong, this initiative falters, I'm going to look after you." Those pieces, social capital needs to be built. And it accrues over time, often quite slowly.

So to the extent that people are turning over more rapidly from firms, which is what we're seeing, every time someone walks out the door, it's not just that individual that is lost, but it's the network of relationships that person had with the people who remain at the organization, and we're losing that. So I think thinking about it as an asset that should be invested in, both as an individual and at the firm level, we obviously think is important and mention ways one might do that.

Hal Weitzman: So is AI necessarily bad for social capital?

John Burrows: No, I don't think so. No, I don't think so. I think we need to look at AI as something that can... It's dual use, right? It can serve good or not so good ends. My sense is if AI makes us more productive in some ways and liberates time, it maybe helps me produce a PowerPoint deck in less time than it might have taken me to do it six months ago, what do I do with that free time? If I spend that time mentoring a teaching assistant or spending more time with a colleague, then I think AI is giving me time to do other value creating things. I think the risk is that the AI frees up that time and then that time isn't used for those social-

Hal Weitzman: It's used to create more PowerPoints.

John Burrows: Exactly. Death by PowerPoint.

Hal Weitzman: Or AI slop, just there's a lot of just more stuff.

John Burrows: Yeah. And I think the other thing we've got to think about is there's two professors at Wharton whose name, they're both economists at Wharton, they've coined this term cognitive surrender. And the idea is, as we all lean more and more on AI, a very real risk is we become less adept at those things we used to be quite adept at. And you would see this with when Excel came along, or the calculator, suddenly doing long division by hand is like a dying art.

AI now can do everything, right? Many of the things. So suddenly simultaneously we're all at risk of becoming perhaps unable to do the things we used to do pre-AI. But I think the cognitive surrender actually is the less worrisome problem. Seth and I talk about relational surrender. I think it's the idea that we're literally going to lose these relationships between and among people. Again, not because I don't want to talk to you how, Hal, but because it's quicker and easier to ask Claude or Grok or ChatGPT. So I think this relational surrender is a term I think needs more attention.

Hal Weitzman: John Burrows, in the first half, we talked about your book, Social Capital at Work, and how it relates to the rollout of AI. There's a trade off there, isn't there? Things are being lost or potentially being lost as we perhaps become more efficient and more productive. I mean, isn't that true of everything, right? It reminds me of slow food, you're the work equivalent of slow food. But it's not necessarily slow. It's about preserving what we have.

You talked about relational surrender there. And there's this great line in the book where you say, you're not worried about machines replacing us, but that we replace each other with machines. We start to have, there's all this psychology now documenting this, our relationships, or young people are much more comfortable talking to a bot than they are talking to another person.

John Burrows: No, I think it's a sad state of affairs. I think AIs of today aren't an entity with which we can build social capital. We can't build trust and rapport. There's no true empathy there. There's no skin in the game for them. But they can simulate a relationship.

Hal Weitzman: I mean, there is a comfort that people get from people who don't want to go to therapy, or can't afford therapy, or don't have access to it and they can get it. So there is some benefit to the individual perhaps.

John Burrows: No, I agree. I agree. But I think at scale, this starts to become a problem. Again, we talked earlier about people retreating into their silos. I only hang out with people like me. I only talk to other marketers. You only talk to other people in accounting. We retreat to our own little bubbles, that could be ideological bubbles, but I think at least we're dealing with other people.

I think many of us now spend an awful lot of our working day not interacting with others, but instead interacting with one or more super intelligences. And I think this is something we should all be concerned about. I'm certainly not anti-AI. I don't think we should shut AI research down. I don't think we should all stop it being used in the workplace at all. I think the gains that can be had in terms of productivity are real, but I think we need to be more conscious of the downsides and aspire to mitigate some of those.

Hal Weitzman: I mean, there's a little bit of a perfect storm situation, isn't there? Given that the move to remote work, to more remote work, the AI thing, the combination of that could be pretty damaging.

John Burrows: Yeah. I mean, look, I think the move to remote work, the conversations were still happening, but you as a junior person maybe didn't hear all of them. Because maybe there was the Zoom meeting, and you heard what occurred in the Zoom meeting, but then a couple of the guys who were on that call happened to be closer friends, and they had a private meeting before and/or after the meeting. But the meetings were still going on. It's just it was harder as someone in remote to get access to that, to learn that stuff.

AI obliterates, has the potential to obliterate, that interaction even occurring. Think of all the things that we used to have to collaborate with others in order to produce various memos, products, Excel models, and such. Now, I can spend days having an AI produce this without talking to another human being. So that interaction is lost. But what's the raw material for our juniors to learn from? I'm going to give them my Claude Cowork prompt list. We need to think about that and go have more conversations.

Hal Weitzman: And presumably it affects, I mean, you're indicating this kind of learning and apprenticeship thing. Presumably that's also about people's confidence and abilities that part of this, the kind of cognitive surrender, is a surrender of skills. Like you say, I can't do long division. I don't need to. I can't create a PowerPoint. I don't need to. I can't do whatever calculations or I can't pitch to a customer, but I don't need to. It can do everything for me. So what were you going to say? Please respond.

John Burrows: No, no. I mean, the response to that, there are a multitude, but one I would have is this, which is...

Hal Weitzman: To clarify, I'm not saying it's a good thing. I'm just looking, this is one trajectory that is feasible.

John Burrows: Yeah. I think we need to recognize that part of work rightfully is building relationships. I think I found this earlier in my career, and I hear this from people I speak to often who are on the more junior side, because at some point you figure out that your worry is not grounded. But you talk to 20 somethings, people in the early 30s, they feel apologetic at times for networking. It feels like this grubby, dirty thing. There's this sense in which, what is my edge? I'm good at talking to other people, as if that's something to apologize for. And I'm going to argue-

Hal Weitzman: The relationship banker on Wall Street.

John Burrows: Yes. What do they do? They just know people, right?

Hal Weitzman: Yeah, just lunches.

John Burrows: They just have lunches. But the point is, that work is real work. It always has been. And more enlightened firms recognize it as such and reward those people. We look for people who have evidence of being able to overcome organizational boundaries, build relationships with other people, work with others, different industries, different cultures, and so on. That's an immensely valuable skill. But that was true before the rise of AI. Now, you might reasonably argue that becomes the sole differentiator. So maybe we should be worrying less about the cognitive surrender issue, though I think that's a valid concern, but I'm much more concerned about the relational surrender, the fact that we're-

Hal Weitzman: So let's turn. So the answer is the three martini lunch will come back. That sounds great.

John Burrows: Let's hope so.

Hal Weitzman: Let's go immediately. Let's talk about, as I said at the beginning, there are so many great stories. I want to make sure we get a couple of them in before we have to finish. One is about the role of warrant officers. You draw this analogy with the military. Tell us about that.

John Burrows: Yeah. So warrant officers in the military have this interesting position in that they're neither part of the enlisted men hierarchy, nor are they part of the officer core. They kind of sit in this awkward middle. They're typically deep domain specialists, but they're also very much loyal to one another as sort of fellow brothers and sisters of the sort of warrant officer cadre.

And what this means is these are people who when, for instance, an aircraft carrier in the Pacific is having a problem with something, some deep technical problem with a nuclear reactor or some other challenge, and the expert who can solve this problem isn't in the Pacific fleet's chain of command. What the warrant officer core is able to do is, because of this sort of mutual sense of obligation to one another, is that warrant officer on the Roosevelt in the Pacific can put a call into a colleague of his in a different command who may be in the Gulf on a different carrier, and can expect reasonably that person will drop what they're doing and solve this problem rather than saying, "That's not my problem because I'm in this command hierarchy and you're in a different command hierarchy."

And I think this idea of the warrant officer is something we should take more seriously in the corporate world. It's not that these people don't exist, these important people who connect across siloed parts of a organization exist. Ron Burt named them. We call them brokers. But very often firms don't know who their brokers are. The Navy knows who its warrant officers are. And I think we need to do a better job organizationally of identifying who these brokers are because these people are the glue, the connective tissue that keeps organizations-

Hal Weitzman: It's different from a general all-purpose project manager because it's someone who knows the organization really well in all its aspects, right?

John Burrows: Yes. And I think a key thing, and this is credit to Ron Burt for establishing this, is there's this sense of knowing different languages. The sense is you think about languages can mark boundaries between spheres of knowledge, but also between organizations. So for instance, let's say I wanted to break into high finance, investment banking. There's a sense in which I can be seen as an insider or outsider by whether I know to refer to discounted cash flow analyses, or I'm a real insider and I say DCFs. That simple mark is like, oh, he's one of us or he's not.

In addition, by learning the language of multiple spheres of knowledge, I can translate and move tacit knowledge across those organizational boundaries. And that's the real value of a warrant officer, or a broker in an organization, is being able to tell the enterprise story and take what might be failure in one part of the business, reframe it and restructure it as success over here, or say, "We're solving this problem in accounting and realize that our friends in engineering have kind of solved an almost identical problem, but they call it something different. But most people in accounting don't speak the language of engineering, so we're oblivious to the existence of the solution."

Brokers are people who have accumulated merit badges. They do all the side quests. They speak the various languages. And I would argue there's immense value to an MBA. What does an MBA provide? It provides you the language of multiple domains of business, which in turn enables you to tell the enterprise story of how businesses work.

Hal Weitzman: Okay. All right. Just one other quick story I want to get in, which is the story about espresso machines and how our physical environment can shape our social capital.

John Burrows: Yeah. So people often ask you, "What can I do? What are some of the steps we can take?" And you can re-architect buildings, and Steve Jobs has done so multiple times, or you can make more modest investments to shape and structure the social infrastructure of an organization.

There's a lovely story about the espresso machines that comes from a guy who's now at GE Vernova, but was at Shell earlier in his career. And one of his initial jobs was to come in and run a research lab where the engineers who were creating these sort of tests and experiments were creating them and handing them off to the technicians and then left the technicians alone to go run these tests. And these two groups of people were literally situated opposite ends of the building and didn't really care for one another. There was a sense in which we're different people. We're all highfalutin engineers and we're the workers doing the real work, not just talking about doing the real work.

And what he needed to do is get these people to interact. They needed to be talking about what was working, what was not working, and iterating. He had a modest budget. And one thing that was implemented was they put in six espresso machines. And this was 20 years ago. This was a long time ago. These were 5,000 euro machines. These were serious machines.

And one of them broke quite early. And what he noticed is the people really wanted their coffee. But it happened that the machine broke that was near the technicians, and suddenly the technician was like, "Well, we got to go over to the engineering end of the building to get our fancy coffee." What happened is suddenly they're interacting. Suddenly while they're making their espressos and having their fancy cappuccinos and so on, they're talking to one another about their experiments, what's working, what's not working.

And he suddenly realized he could strategically break the espresso machines, not actually break them because they were expensive, but could break the machines and make it look like they're out of order to literally affect the flow of foot traffic throughout the corporate campus, which I think was a really lovely story.

At the other extreme, you've got Steve Jobs both at Pixar and Apple when he built the spaceship flying saucer looking building, re-architected headquarters in order to further the creation of social capital very intentionally. And Joel Podolny, who was actually hired by Steve Jobs, Joel Podolny was the dean of Yale School of Management when he was poached by Steve Jobs to stand up Apple University and ultimately became Apple's head of HR. He was a network theorist. He was a social capital guy. He was a peer of Ron Burt's.

So Joel Podolny tells us a story of how the reason that spaceship, that flying saucer looking headquarters, is the way it is because Steve Jobs wanted to maximize serendipity, wanted to maximize the interactions between people. How do you do that? You don't have straight corridors. You have a long circular corridor. And at Pixar, he wanted to structure it so everyone had to converge into a central atrium and force them to interact, but then could retreat to small, cozy private spaces. So the thought was both to increase open networks and build weak ties, the serendipitous accidents that occur in the central open atrium, but then I can retreat back to a private space with two or three intimates on my team member or future collaborators to talk more privately.

Hal Weitzman: Even in the age of AI, people will still be drinking coffee.

John Burrows: I hope so.

Hal Weitzman: So that was the work. John Burrows, thank you so much for coming on the podcast to talk to us about your book, Social Capital at Work: Building the Hidden Asset that Drives Trust, Engagement, and Performance.

John Burrows: Thanks for having me, Hal. It's been a delight.