How AI Can Reveal the Neural Secrets Behind Great Leadership

MOLLY WOOD: Today, I’m talking to Michael Platt, Director of the Wharton Neuroscience Initiative, and Professor of Marketing, Neuroscience, and Psychology at the University of Pennsylvania. He researches how our brains work in relation to communication, decision making, group intelligence, and team building. He’s also written a book on the neuroscience of leadership. He believes that AI-driven insights from neuroscience can transform your business, and he’s here to explain how smart companies are already taking advantage of this potential. Michael, thanks so much for coming on the show.  

MICHAEL PLATT: Well, thank you, Molly. It’s such a pleasure to be here.  

MOLLY WOOD: So I’d like to start by asking you about the neuroscience of leadership. What would you say is a fundamental insight from your research about the secret to being a great leader?  

MICHAEL PLATT: There are certain capacities that seem to be super important, and one of the most important, I think, in my research that I’ve found is the ability to connect with others and to create context and to create spaces where people can connect with each other, and that is such an important part of who we are as a species. It’s a critical driver of productivity, and it is the foundation of wellbeing. And when we think about all those things—productivity, creativity, innovation, harmony, team dynamics—they all really start at the top. And that’s where we often see the failures of leadership. And so if you don’t create a harmonious culture, if you don’t lead with integrity, if you’re not a good communicator, you know, then things start to fall apart. 

MOLLY WOOD: This has been, I think, one of the themes of this season and even last season, this idea of the importance of humanity and empathy. And one of the key leadership insights in your book is, in fact, that humans are fundamentally social creatures. And that a huge part of leadership is facilitating these strong social connections. How do the best leaders do that?  

MICHAEL PLATT: I think people are beginning to be more mindful about it because in the past, leadership was thought to be, you know, you’re kind of removed from the people that you actually lead. And, in fact, being present, devoting your attention, it’s actually kind of where things start because our brains are actually quite limited in their capacity, unlike AI. And so where we attend is telling other brains exactly what’s important, and if you’re not paying attention to the people around you, their brains know it. So I think what’s really important here is for leaders to recognize how important connection is for effective team dynamics, for creating a culture where people show empathy, where they can work harmoniously and agile, you know, they’re agile and they can navigate through difficult times, because we also know that that social support is critical for change management. So, how do you do that? 

MOLLY WOOD: No big deal. How do you do that? [Laughs] 

MICHAEL PLATT: Well, there’s a lot of ideas out there, but it often feels like, how do you know which ones to use? And what’s really going to move the needle? And that’s where I think neuroscience can help, because it can go beyond the surface, go beyond the sort of self-reports and surveys, which people aren’t always honest on or they’re not always as self-aware or as in touch or, you know, sometimes you want to hide things, or it’s biased, et cetera. We can surface what’s going on underneath the hood and that can tell us something about how effective a particular strategy, a particular intervention, a particular context might be for enabling people to form the strongest connections that they can. 

MOLLY WOOD: Can you give us some examples of strategies that work and maybe strategies that, you know, don’t work as well as we would like to believe they do? I’m thinking, do we have to keep doing trust falls at off-sites? What are we talking here? [Laughter] 

MICHAEL PLATT: We have not looked at the neuroscience of trust falls. [Laughter] But one that we’ve really been working hard on, and I think shows strong effectiveness, is creating the space for authentic conversations. We work from a set of structured conversations that was developed about 15 years ago out of UCLA, something called “Fast Friends.” There are other versions of this. It’s designed to take people from a space of almost not knowing each other. In fact, we work with people who are, who we pull off the street, you know, random students, and it compresses the process of becoming close friends, which might take months, and it compresses it into like an hour. So it starts with a very chitchat-like icebreaker kind of questions. And then moves rapidly into things that are quite deep. But when you’re authentic, when you give your attention, as I mentioned before, and when you listen when somebody else is speaking, this is what is the foundation of trust. And when we make eye contact and we listen and we speak to each other authentically, this engages circuitry in our brains, which is the foundation of our social connections. It’s something called the social brain network. It manages our moment-to-moment interactions with others, as well as our long-term relationships that are stored in there. And it doesn’t really get activated unless you’re really paying attention and listening. But when two of these social brain networks get going, they go into resonance, what we call brain synchrony. And when I think about it, it still feels like witchcraft, black magic, to me. There’s this observation that’s been made many, many times in humans, in social animals, that when you and I are having this conversation, that our brains would begin to resonate, which means that the same patterns of activity in my brain would be observed in your brain, and that indicates that we are kind of seeing the world the same way. We have similar emotions, similar thoughts, and, remarkably, that percolates down to our bodies too. So our hearts begin to beat together. You can observe this. I do this as demos all the time.  

MOLLY WOOD: Wild.  

MICHAEL PLATT: And you can run it in reverse. So you can, if you move together, you breathe together, you can synchronize your brain activity. When you make eye contact, your brains actually synchronize. This is what evolution endowed us with over the course of many tens of millions of years, is this capacity to get in sync with others. And when we do that, that’s what we call a biomarker, a biological marker, that predicts increased trust, better communication, better teamwork, you name it. If you’re in a better relationship, if you’re, you know, because it happens between partners, a better relationship with your child, it’s sort of the cornerstone of our social connections. And one of the things that we found, this paper that’s coming out soon, is that one concern would be that, well, you have a team of people, you have a committee at work. You’re all synchronized. So as that group you just kind of like forget innovation and creativity, you’re just stuck. And what we found is exactly the opposite. Teams, committees—these were actually committees who were evaluating candidates for a job—when they were in sync with each other this allowed for people to feel psychological safety. We think that it’s actually a biological marker of psychological safety, the first one identified. And when they have psychological safety, they feel confident enough to volunteer information for full, robust discussion, which otherwise wouldn’t come out.  

MOLLY WOOD: And what role does AI have to play in fostering these connections that are so fundamentally important to leadership?  

MICHAEL PLATT: AI, on the one hand, is similar to the human eye. So, organic or natural intelligence. But there are a lot of things that are different. All the buzz is about large language models. They are conversational, right? And so, they’re very responsive to us, and studies show that actually they can be more persuasive than a real live human. Or they can be, you know, like an AI therapist or AI doctor is rated as warmer, more empathetic, people would prefer the telemedicine from the AI. And going back to the science of synchrony, I think there’s a really interesting insight there, which is that—and maybe it’s already being done, I don’t know—is to build synchrony into the expressions of the AI that would be kind of personalized. I mean, we know that people—there are other studies of just human beings talking to each other and people have close relationships—tend to show a higher frequency of saying similar things. So I think that there’s a way to build that into these large language models so that people inherently trust them more. I mean, there’s dangers there, too, but I think that’s kind of what we should be thinking about is, not so much how we prompt the AIs, but how they kind of prompt us in a conversation. And to take that further, could we build that into the way that AIs prompt a team of people to coax along a conversation like a facilitator so that the team gets more in sync with each other and potentially more in sync with the AI.  

MOLLY WOOD: Right. So there are parts of what you’re studying that can be replicated verbally, interactively.  

MICHAEL PLATT: Correct.  

MOLLY WOOD: Separate from these kind of physiological responses.  

MICHAEL PLATT: Right. So we do the fundamental hardcore neuroscience that might be in the brain, it might be cellular, molecular, et cetera. Brain imaging, you know, pretty lights on a brain, brain scans on people. But we’re always looking for across the continuum of less and less obtrusive measures, more scalable measures. At the very end, even what we say, how we say it, how we move, are potential indicators of our internal states, of our mental and emotional states. I think there is good evidence that, for example, you see synchrony in these things and that when you leverage, say, moving together, that’s a really good example. So if you want to get in sync with another person or with a team of people, the idea is you try to mirror somebody else’s movements as closely as possible. And data shows that when you do that, you actually synchronize your physiology. From an anthropological perspective, every society on the face of the planet has rhythmic group activities that they engage in. Dancing, singing, chanting, drumming. And they seem perfectly designed to build synchrony amongst people. It seems to me that should be built into AIs to recognize and leverage and enhance synchrony with a user, but also amongst groups of users.  

MOLLY WOOD: I really want pretty lights on a brain on a T-shirt, by the way, that’s a great, it’s just a T-shirt design waiting to happen. [Laughter] So I’m hoping to dig into how the brain makes decisions and how AI can facilitate and improve that. I know you have talked about AI’s potential impact on business in your field. And you’ve noted that AI, like humans, can be really good at making predictions, even though it goes about it very differently than our brains do. Can you talk a little more about that?   

MICHAEL PLATT: Think about the human brain as one big prediction engine. That’s how it’s designed. It’s constantly making forecasts based on previous experience and even the experience of our ancestors. So our brains are making forecasts and then checking, was that forecast met or did things turn out better or did they turn out worse? And then that error signal is kind of fed back in so everything improves, hopefully. Pattern recognition and pattern generation are sort of a slightly different version of that. And, you know, when you think about AI, both of those two components are there as well. So there’s the prediction error–generated learning, as you see in machine learning, algorithms, and then the chatbots are more oriented around pattern recognition and pattern generation, which is a little, slightly different process. And so, when I think about the comparisons—obviously they’re embodied very differently, AI and this thing here, this meat and fat between my ears—but the resemblance computationally is becoming really strong. And that’s really fascinating to me, because for most of the 20th century the argument was that you couldn’t build something smart like a human brain. Turns out you can, just if you have a lot of computing power. The major difference right now is the efficiency. These big AI models require vast server farms, data centers, chewing up energy. This thing in here runs on the equivalent of a 12-watt light bulb. [Laughs] 

MOLLY WOOD: And maybe a little Red Bull here and there.  

MICHAEL PLATT: Well, for sure, some Red Bull... But that’s a big difference. And so that’s, I think that’s one of the things that’s computationally distinct, is that there’s a lot of efficiencies built into the way the human brain works. Like, it throws out a lot of information. That’s one of the most important things it does. It doesn’t encode everything, whereas these AIs are kind of encoding everything.   

MOLLY WOOD: I also wonder how the reverse might be true, how AI could help improve our predictions, because our predictions are so likely to be inflected with what we maybe think we know.  

MICHAEL PLATT: For sure. [Laughter] I mean, there’s so many, well, yeah, exactly. So what I’ve been saying in terms of the past being baked into our brains, you know, our brains were not designed for this environment. They were designed for navigating in small groups of people in the savanna or the sort of grasslands of Europe or Asia, and so there are all of these anomalies. We don’t think like computers, at all. We’re afraid of loss, even when there could be a good bet there for us. And so we require extra payments. We don’t give up things that, even if they’re worthless, you hand them to us and now we don’t want to let go of them—the endowment effect. And the list goes on and on and on.  

MOLLY WOOD: So this historical knowledge, this sense of what we think we know, can potentially maybe be corrected by AI’s ability to make predictions based on more reliable data sets.  

MICHAEL PLATT: More and more and more data. So the more data you have, the lower the uncertainty in general. And our brains are are limited. Our brains are fallible. Our brains throw out lots of information. We orient toward things that stick out, that stand out—the shiny bright moving things—and we might be missing something. Getting back to social connection, which is awesome, or doing things that require more than one individual. But it also means that there’s a lot of social contagion, that there’s a tendency to believe what other people believe. We also are attracted to and value the behavior of individuals who are of higher status than us. All of those things shape our behavior in ways that in today’s environment are probably not always adaptive, and that’s where AI could certainly be a corrective. Insofar as we can find a way to, you know, this would be the AI-assisted nudge, because we’re not going to rewrite the rules that our brains live by. Let’s just kind of take that as a given, and we’re not going to use AI to probably, to stimulate our brains to change those rules or anything like that. So we’re going to have to use some kind of nudges, and so maybe AI can nudge us. And that’s probably true. I mean, we’ve already seen that to some degree working in apps that use principles of behavioral science and nudging to help people lose weight and to eat healthier diets, you know, to avoid screen time and things like that. 

MOLLY WOOD: When you have been asked about the business impact of neuroscience and AI, you point to this concept of neurodata, and you say, if companies are not collecting neurodata, they are leaving high-quality data on the floor. First of all, what is that?  

MICHAEL PLATT: There’s a vast array of different kinds of neurologically relevant data. I don’t imagine that in most companies they’re going to develop a wearable version of an MRI machine to put on every employee’s head. But, we have, for example, and I know that there are companies that have been doing this, using wearable brainwave-sensing devices. A headband, or a hat, or some companies have developed headphones, or earphones, and can get some kind of data from that. That can be really informative. Because it, again, it’s giving us access to something that is covert, that’s hidden, and that may be at variance with what people are saying. And, you know, a lot of my book, a lot of the work that we do, is showing that there’s often a major disconnect between what people say and what their brains tell us. Classic examples are in the case of unconscious bias or implicit bias. You show somebody a video of a person in pain, and the person can be from their ethnic group or different ethnic group, and they say they feel the same amount of pain for each person, but their brain empathy response is totally different, and is tribal. And that’s why unconscious bias is so pernicious and difficult to overcome because you think you don’t have it, or you say you don’t have it, but it’s there. So, any purchase on that—so, even these brainwave signals can tell us something about how focused you are versus how distracted you might be, or how frustrated you are versus how calm you might be, or how forward-oriented. Like, I really like being here and talking to you, versus my brain is out the door. And there are other signals that are easier to acquire, like facial expressions, tone of voice, that tell us something, but it’s less, right? So there’s, as we get more removed from directly accessing or indirectly accessing the brain, the information content is lower. So heart rate, like we’re getting from wrist watch devices, wrist-worn devices. Yeah, it tells you something about your activity level and your fitness. It tells you a little bit about how excited you are in the moment, how roused you are. And that can be a predictor of subsequent behavior. We’ve shown that heart rate synchrony amongst people is a good predictor of their team dynamics, for example. And that’s pretty—it’s easy, right? So that’s easy to get. And I think all of this is going to get easier to do with just improvements and miniaturization in tech. The more we collect this data in real-world environments, real work environments of varying complexity, the AI will allow us to make more sense of that data and connect that to more different kinds of outcomes, which is great for healthcare, it’s great for a lot of different things. The biggest issue here is the, you know, the ethical, legal, and societal implications. [Laughter] So, it’s like, of course we should be worried about this, it’s bad enough—   

MOLLY WOOD: I’m not saying I’ve been thinking that this whole time, but I have been wondering how you—  

MICHAEL PLATT: Everything we do in this space is really, you know, it raises all these serious issues. 

MOLLY WOOD: Right.  

MICHAEL PLATT: Honestly, it’s like the other issues are, they’re really bedeviling. Because it’s all about, like, how are you going to use it? And that’s sort of the Peter Parker principle that I try to abide by, but, you know, we have to make that decision as a society, because with all this power comes great responsibility. 

MOLLY WOOD: I want to go back to neurodata for a second on that point, because it does seem, obviously, there are a lot of real questions to be raised about how that data is gathered and who opts in to offering that data. But given the gap that you’ve described between what we feel and think and what is actually happening in our brains, if we are to train AI and give AI data sets that are accurate, it sounds like you’re saying neurodata is crucial there. And then you do start to get to, for example, more honest, more equitable AI because it has better information.  

MICHAEL PLATT: I think that’s a really great point, and it’s one that has surfaced many times in the courses that I teach, which are essentially neuroentrepreneurship courses, and I teach MBA students, for example, a little bit about neuroscience, and then they run with it to develop applications and pitch companies. And a lot of what I’ve seen there is a way to de-bias many of the things that are currently so biased, whether that’s de-biasing, policing, or matchmaking. 

MOLLY WOOD: Let’s bring this back to, I mean, I also want to take all the neuroentrepreneurship classes immediately. How do we continue to apply this in kind of a business leadership context? How does this data help companies understand their customers, in addition to their employees, build better products, all of that?

MICHAEL PLATT: The pitch for understanding your customers is a really easy one. So, the sort of neuromarketing and brand strategy, the lift is demonstrable. It is getting way more sophisticated because you can basically, not only do A-B testing on steroids and know exactly how to optimize ads, that’s been known for a while, but you can target those ads to the people that they most effectively move, that they reach on an emotional level, and are almost guaranteed to move them to the decision that you want. 

MOLLY WOOD: Yeah.  

MICHAEL PLATT: Given that, you know, we’ve been in fact using AI to do a much deeper dive on the question, for example, of what is the meaning of a brand in the brain of a customer, and how does that relate to the equity of that brand, the value of that brand? I mean, companies spend so much money to create a brand and use that to inform and acquire customers, but you want to keep those customers, you want to keep them for a lifetime. And some companies are really good at that. And what we found is if you just look for those pretty lights on a brain, but not like how much light is there, but where the lights are and which ones are on and which ones are off, we can precisely predict sales. We can precisely predict sensitivity to changes in prices and things like that. And so what we do is we use AI to kind of decode that into psychological terms that we understand, based on the tens of thousands of lights-on-a-brain studies that have been published in the past. Then that gives companies, that gives creative and marketing, a handle on what to work with. Like, it turns out that this brand works because it stimulates nostalgia in the brain, and this brand works because it stimulates social connection, and this, you know, and then if you’re going to come into that space and you’re going to position your brand, you can leverage those tools to do so in a much more effective and efficient way. Look, in business, there’s always a brain, at least right now. [Laughs] It’s not all AI-driven, and often more than one brain is involved. And so if we can understand those brains better, then we can make that whole process work more efficiently, more effectively, and more humanely.  

MOLLY WOOD: Michael Platt, author, professor, and researcher. Thank you so much for a fascinating conversation.  

MICHAEL PLATT: Well, thank you, Molly.  

[Music] 

MOLLY WOOD: Thank you again to Michael Platt, Director of the Wharton Neuroscience Initiative and Professor of Marketing, Neuroscience, and Psychology at the University of Pennsylvania. Please subscribe to WorkLab and check back for the next episode, where I’ll be speaking with Charles Lamanna, Corporate Vice President of Business Apps and Platforms at Microsoft, about how AI will lead to nothing less than complete business reinvention. If you’ve got a question or a comment, please drop us an email at worklab@microsoft.com, and check out Microsoft’s Work Trend Indexes and the WorkLab digital publication, where you’ll find all of our episodes, along with thoughtful stories that explore how business leaders are thriving in today’s new world of work. You can find all of it at Microsoft.com/WorkLab. As for this podcast, please rate us, review us, and follow us wherever you listen. It helps us out a ton. The WorkLab podcast is a place for experts to share their insights and opinions. As students of the future of work, Microsoft values inputs from a diverse set of voices. That said, the opinions and findings of our guests are their own, and they may not necessarily reflect Microsoft’s own research or positions. WorkLab is produced by Microsoft with Godfrey Dadich Partners and Reasonable Volume. I’m your host, Molly Wood. Sharon Kallander and Matthew Duncan produced this podcast. Jessica Voelker is the WorkLab editor.