Available October 29, 2026

Episode 039: Who Leads When AI Thinks? Judgment, Ambiguity, and the Information That Never Reaches the Decision (Part 1)

DURATION: 62 minutes
HOST: Kevin Novak
GUEST: Adrian Wolfberg, Ph.D., Founder and CEO of Organizational Insight Consulting and Author of Who Leads When AI Thinks?

FORMAT: Guest Conversation, Part 1 of 2

SHOW: The Human Factor Podcast

COLD OPEN

Kevin Novak: Here’s something that happens in every organization every day. And well, almost nobody sees it while it’s happening. A piece of information starts at the edge. Someone on a sales call hears a customer hesitating. An analyst notices a number that doesn’t fit. A nurse sees a pattern on the floor that the chart simply doesn’t show. That information starts moving towards the person who decides. At every step, someone summarizes it, someone tidies it up. Someone drops the caveats because the caveats make the slide really messy. By the time it reaches the top of the building, all the evidence is gone. Only the conclusion is left. The decision maker gets a really clear answer, but no one has an idea how much doubt was stripped out along the way. In 1958, James March and Herbert Simon gave that a name. Uncertainty absorption. It’s one of the oldest findings in organizational research. And we have done almost nothing about it in nearly 70 years. Now put a machine in that chain. A system that produces confident, well-written conclusions at any hour of the day. No memory of where its evidence came from. No way to tell you what it left out. The telephone game just got a new player, and it never gets tired. So the question isn’t whether AI can think. The question is who leads when it does? And whether the human at the end of the chain still knows how to ask where an answer came from. I’m Kevin Novak, CEO of 2040 Digital.

INTRODUCTION

Kevin Novak: Professor at the University of Maryland and author of the book The Truth About Transformation, Leading in the Age of AI, Uncertainty, and Human Complexity, along with the Ideas and Innovations weekly newsletter. Welcome to the Human Factor Podcast, the show that explores the intersection of humanity, technology, and transformation. Along with the psychology behind transformation success. This is season three. If you’ve been with us since the opening episode, you know what this season is about. Practice, not an initiative, not another workshop, but what it takes to build the human factor into how an organization works, so that it survives after the consultant leaves. The first thread running through the whole season is the dynamic between artificial intelligence and human beings. I keep calling it the clearest case study of the human factor that we have ever had. And it’s unfolding right now in real time. Today’s guest arrived at that thread from a direction almost nobody else in this conversation can claim. Adrian Wolfberg spent nearly 40 years as an intelligence officer. He started as a naval flight officer, flying carrier based reconnaissance missions, and out there, teamwork and confidence under uncertainty. Were not concepts just in a leadership book. They were the difference between a good day and a very bad one. He went on to the Pentagon, briefing senior leaders up through one-on-ones with the Chairman of the Joint Chiefs of Staff. That is where he learned something, the thing his career has been built on ever since. Knowledge that can’t cross the boundary between the person who produces it and the person who decides with it might as well not exist. That problem became his doctorate at Case Western Reserve. It became the knowledge lab he created and led at the Defense Intelligence Agency, which was an internal consulting capability that helped people solve their organizational problems. It survived two changes of command, which anyone who has ever worked inside a large institution will recognize as its own kind of achievement. His research since then has followed the same pattern in the business with elite athletes and in the high-pressure healthcare teams. Today, Adrian is the founder and CEO of Organizational Insight Consulting, and he’s the author of Who Leads When AI Thinks? How leaders frame decisions and exercise judgment. The book’s argument. Is one I want you to hold on to for this whole hour. The human element becomes so much more important in the age of AI, not less, and I can’t express that enough. Judgment, context, empathy, and how you frame a problem are the competencies that decide what the technology is worth. And a leader has to understand the problem before they ever touch the tool. He also has a second book in editing right now, and we’re gonna get to it names something every organization feels, and none of them have a word for. Adrian, welcome to the Human Factor Podcast.

SEGMENT 1: FROM THE CARRIER DECK TO THE KNOWLEDGE LAB

Adrian Wolfberg: Thank you, Kevin. Thank you for inviting me. Glad to be here.

Kevin Novak: I’m excited for this episode. So, Adrian, I want to start with the arc of this season. So the audience needs to understand where this comes from before we get anywhere near the technology. There’s a thread that runs through this show since the season two finale, and it’s resistance is information. When people push back on a change, that pushback is a diagnostic signal in many ways. It tells you where trust is thin, where people feel unsafe, where somebody’s identity is on the line. And most leaders treat it as a problem to be managed rather than a message to be read. So you told me that running the knowledge lab was where that idea became crystal clear for you. The resistance told you exactly where the negotiations needed to be focused on. But you’ve been building towards it for a long time before that. Flying reconnaissance off aircraft carriers, then sitting in rooms in the Pentagon watching knowledge fail to reach the people who actually needed it. So walk me through that arc. When did you first understand that the obstacle was never the information itself?

Adrian Wolfberg: Yeah, I guess at first I guess it first happened certainly flying off aircraft carriers provided some kind of background, but I can think about a specific incident in the Pentagon. So after doing after flying off aircraft carriers for about four years, then I went to the Pentagon and there was a there was an operation that I was involved in the preparation. The up the operation was going to take place overseas where US forces would be in harm’s way and my role was to not only be in the preparation part of the operation, but during the operation monitor how things went in case things went sideways. And there was something that I then during the operation, there was something that I noticed based on the information that was coming to me, that something was not expected and that may change the balance between a superiority of American forces and what was perceived as a lack of ability of an adversarial force. And with this information I came to the conclusion that the balance had been changed, that it was more on parity, which was unexpected. And that was the knowledge that came to me through my analysis, and I was prepared then to disseminate that information through a local, because it was monitoring an operation, it was real time. So I was with a bunch of people and there was a leader, maybe a dozen or so people. And the people in this group, including the leader, said, where did you get this knowledge from? And I said, well, I went through this analysis myself and they said, Well, let’s you know show us who told you this. And I said, Well, no, you it’s that’s not like that. It’s I came to this conclusion. And they basically just you know laughed me off. And I don’t remember how it happened, but one of the individuals who I worked with during the planning phase. I may have contacted him and I don’t remember exactly. But anyway, he came up to me in the midst of this group that we were sitting in the same room and I told him the I told him the rest of the story and he said, You need to come with me right away. So I said, Well, where are we going? And he said, you just follow me. We’re gonna go somewhere. We’re gonna walk, it’s gonna take us about a minute to get there. And he on the way, he said, Now, when we get into this room, you cannot go into all these details that you know that I know, because the person you’re gonna be talking to was never privy to this kind of detailed information. And you have to be able to explain what it is. And what the impact is in a minute or less. So as we were walking, I put my head together to be able to do that, not knowing what the purpose of this meet you know conversation was going to be or who I was going to be talking to. And then went into this room, and it was a small room, and there was like a stage, a small stage, and there was a guy sitting. A big guy sitting in the middle of that stage with a big chair, and lo and behold, it was the most senior military officer in the US, in the United States, the Chairman of the Joint Chiefs of Staff. And I was just standing about three or four feet from him or less, and I went through the spiel in less than a minute, and he then warned the forces who were overseas so that they wouldn’t be surprised by this unexpected activity. And so, that taught me that knowledge by itself is just the beginning and it has to travel. And it has to travel through potential barriers. perhaps there’s a little barrier, perhaps there’s a lot of barriers. In this case there was a lot of barriers. My reputation was on the line, I was laughed at. And there this traveling has to culminate in someone being able to absorb information that they can’t be confused about. They have to know just enough that they have to kind of have the bottom line but to be able to judge your reputation, your th to understand how you got to where you got to and forth. And that was like an eye opening experience that turning knowledge into action was more than just the knowledge itself.

Kevin Novak: So you said something to me on our first call that I’ve been carrying around since. And that’s that organizations push problems up a chain of command, where in your words, they simply don’t exist as problems. And what you just talked through demonstrates that really well. So someone at the top is being asked to solve something that they can’t see. By people who can see it and can’t act on it. So what did the knowledge lab teach you about why organizations do that? And what did it take to get people to solve their problems instead?

Adrian Wolfberg: Well, these I guess what it taught me was that there’s an expectation, and this is not just in the organization that I was in, because later in other contexts I would hear the same thing, and when I did research there was the same thing. But in that particular experience, I could see that there was an expectation by people who were my peers or superiors that someone else was gonna solve the problem and that all that was needed was to communicate the problem up the line so that there needed to be kind of a strategy of who to target, what level leadership needed to be targeted, but it

Kevin Novak: Sorry.

Adrian Wolfberg: Was this idea that you needed to you need to identify a leader at the top who one felt that could solve the problem. And in a sense, those who were experiencing the problem just has to pa just had to pass that on. So it was kind of like action passed, action complete, as if the transmission of the nature of the problem would be received, absorbed, and action would be taken. And that often does not work. And even if there is a reception or an acknowledgement of underst of you know an apparent understanding of what the problem is. There’s no guarantee that any action will occur, either on a timely basis, or even if the action is targeted to the underlying problem as opposed to the symptom. So there’s this, and in the military, I was in the military in the Navy, and I, I saw that well before the knowledge lab. The whole hierarchical construction of organizations that probably goes back to the Industrial Revolution in which there were intentionally levels of control designed to manage subsystems within an organization. And that you know led to people at different levels of the hierarchy having different sorts of goals besides knowledge integration. This would be power resource control things about reputation to reputation up to the those above them, all these kind of other factors that come into play. so that’s what I saw and at the time about a year or so into this knowledge lab experiment which I’m gonna just talk about one of those was over fifty kind of projects or experiments, kind of experiential learning of helping people move through the problem so that they could actually be part of the solution and create the solution. it just it turned out that a new leader came into the organization and had identified through the annual surveys, these are employee surveys that judge temperament and morale and forth. And people had had entered into the open field areas as opposed to the multiple choice questions, issues that were concerned that weren’t addressed in the multiple choice, and a strong theme came out of distrust with leaders in the organization. And that was kind of the extent of it. And this new leader came in and wanted to change that dynamic. So there would be, his goal was to have people feel more positive about leaders. Well then the question was, how do you do that? How do you how do you I mean this is an this was an overwhelming lesson in the knowledge lab is how do you change people’s mindset when they’ve never experienced that mindset in the first place? How do you move from not knowing something to then knowing something? it’s one thing if you’ve had an experience.

Kevin Novak: Well, Adrian, I’d suggest that it’s almost a learned behavior. It’s a training that people go through. It’s a stereotypical expectation that one doesn’t do this and those lower don’t do that in informative ways.

Adrian Wolfberg: Yeah, I mean it’s reinforced by culture, by norms, by policies, by structure, absolutely. And so my team and I figured out a way to achieve his goal. He didn’t have the it’s an implementation issue, an implementation challenge that one would imagine is quite challenging. How do you get people to believe have a positive opinion about leadership when they had a negative opinion about leadership? And we came up with this idea and because he also knew that there was this challenge of as you were ta as you were talking about information going up and making its way past the filters and the barriers to see if it actually reaches the top. So we

Kevin Novak: Thank you.

Adrian Wolfberg: Constructed a face-to-face venue in an auditorium where employees of any rank or seniority could come once a month and the leader would be in the front of the stage and he would ask, Okay, he would just open up and say, first of all, this is not a complaint session and if you do have a concern, you’re going to be part of the solution, part of exploring how to do the solution. And the people in the auditorium or in the broader organization have their the right to join you in exp in exploring how to solve this issue. So what actually would happen is that someone in the audience would explain an issue and he would then say, yeah, I think that’s a worthwhile issue and I authorize you to pursue that. And my role and my team was to facilitate that individual with other individuals to actually go through the hard process of coming out on the other end with a implementation solution and forth. And this was very successful until people who were between the leader and the frontline employee, the second level supervisor, the kind of the executive who was the direct report to the leader, got wind of what was going on and they were upset that people were bypassing the silo structure that they each person was employed in and didn’t make this aware to the particular executive or senior manager. And that was you know potentially embarrassing to the leader who now is aware that there’s a problem that hadn’t been fixed. And that would be a threat to the executive or the senior manager. And the senior so the you know the direct reports or this senior managers told people that they could not come to this session unless they get approval and all that sort of stuff, which was contrary to the whole intent of this leader’s purpose. And I found out about it because people complained to me that this was going on, because people found that this was a breath of fresh air that there was now an opportunity to solve these at each of these problems had been perceived to be intractable. In other words, they had not been either able to be articulated, to be communicated, or if they had been articulated, nothing had been done about it. And these were things that were really bothering people, either in life, work-life issues or the work itself. And the I told the leader this and he communicated to his direct reports and then that cascaded downward that this is not a reflection of them not doing their job. This is just kind of a natural part of the challenge of working in hierarchies. And he wanted them to then attend the meetings and to be willing to participate in giving their views about helping people who may or may not be under working under them to move forward to a problem to a solution, which they did. And the director often the head of the organization often called upon them you know impromptu to give their opinion and forth. so that was an example of where there was resistance and there was you know a strategy then to deal with that resistance. that was like a very obvious example.

SEGMENT 2: UNCERTAINTY, AMBIGUITY, AND THE TIME WE NO LONGER GIVE IT

Kevin Novak: So, with all that, and you’ve well demonstrated that information itself was not the obstacle, right? And it sets up a distinction that I most want the audience to hear from you. And you draw a line that most leaders never draw. And it changes what they should be doing with their mornings. Uncertainty can be resolved by getting more information. You don’t know a number, you go find the number. And ambiguity cannot. Ambiguity is about making meaning from information that will never be complete. And meaning making takes time and it takes diverse perspectives, which is what you are walking through, and it takes deliberate framing of the problem. So to me, Karl Weick is one of my favorite researchers, and it’s one of my favorite territories that he talks through. And Weick spent his career on what he called sense making, the process by which people and organizations construct a reality or a version of reality and then act on the version that they build. And to your point where they’re only you only have so much information, you only have 60 seconds to explain what you’re asking them to solve. So his point was that leaders don’t perceive the world and then respond to it. They make sense of it first. And the sense making is where the decision actually gets made. So now here’s a collision. The AI era compresses exactly the time that meaning making requires. And the answer arrives in seconds. So pressure is to act in seconds. We feel that immediacy. And most often we’re always thinking in the short term. So one of my longstanding mantras at 2040 is that speed kills with uninformed urgency. And I’ve watched more than one organization run itself into a wall at high speed with very confident dashboards on the windshield, right? So make the case to a leader who believes speed is an advantage. What does it look like practically to treat time as a value on the front end of a decision rather than a constraint, which is what is so much the behavior out there.

Adrian Wolfberg: Yeah, no, that’s really good. I totally in agreement with you. yeah, I would distinguish between well, I first would say, yeah, speed is important, but it depends. And there is this decision speed, and then there’s this decision quality over time. And you can’t well so treating time on the front end as a value, I think in the era of AI, is where you structure your approach to solving an issue such that you actually will have a competitive advantage over people who just leaders who just take speed as a to make a decision really quickly just to move things on and to get things going, kind of a thing. And what I would recommend and what I talk about in my book, Who Leads When AI Thinks, is that the front end of an issue that you’re facing has to be devoted to taking the time. You have to make the time to thoroughly understand the problem. Now that should always have been the case, but in the era of AI, this accentuates the need for it. And what that means is to make sure that there’s enough perspectives to understand the breadth and the depth of what the problem is really about, what second and third order forces are operating on the organization, whether it’s internal, external. So that on the back end, as you move through working through the relationship between AI and humans, which is this conversation we can go into next or later, that you’re optimizing in a sense that knowledge integration between these two intelligences. And you don’t fall you’re more likely to come out with a workable solution as opposed to this decision speed, which is I think what most people are operating in, to just get action started and then find halfway down the way since the world is changing and just you had a limited view in the first place, you need to do medium or major kind of adjustments, reallocate resources, take additional time, and you’re just kind of wasting resources.

Kevin Novak: So, Adrian, what happens downstream as you hit on the front end? And it’s when the front end work doesn’t actually happen. And in my experience, it doesn’t show up as a bad decision. It shows up, as you mentioned, six months, six months later, three months later. So three teams wind up solving three different problems that all have the same name. How what’s the solution out of that?

Adrian Wolfberg: Well, in this front end, it’s all about the front end and what’s going on in the front end and as time moves on. so on the front end, there’s not only the awareness of the nature of the problem, but there’s also the next step on the front end, which is the consideration of what’s important in the context that the organization is working in. There’s this kind of universe of things that could be affecting or either actually or potentially affecting achieving whatever the goal is to solve this problem. there are going to be discussions and decisions made about which force or which factor in the world is worthy of being considered and why it’s being considered and what the trade-offs are and what the what you what the risks are and what you’re willing to live with and things that you need to mitigate. All these kinds of variables that come into this equation are discussed in the front end. And if that sort of knowledge, not only the knowledge of the nature of the problem, but the knowledge of how the problem has been framed, does not travel with the resources that have been assigned, obviously that may or may not have to think about how that would apply to a large language model or some other form of AI, but certainly to humans. If that doesn’t travel to humans, then they don’t have that information a and either they will have to pause and go back up the chain of command for clarification about what needs to be done, which involves some kind of both time delay to do that and translation distortion going up, going back to the problem of what is perceived as a problem at one place is not perceived as a problem of another or it’s perceived differently. So that creates lots of inefficiencies and possibly ineffectiveness when this background knowledge on the front end is not does not travel. And then it and then if there are multiple groups that are working on this, either in parallel or in series, and the that this knowledge about this kind of meta-knowledge about why and why we’re doing things and what’s important and et cetera, isn’t transparent and doesn’t travel with. It just kind of cascades through the system. Things that only become doable based on the viewpoint of where people sit in the organization and what they think is important or what the managers above them think is important that they should work on. So you get at the top you get someone who feels like they’ve resolved the frame of this problem and the all the reasons why information is in the frame, all these decision factors, the assumptions made, the risk being associated with it, all this, the mitigation strategies. I if that doesn’t cascade throughout, then you have this kind of inverse uncertainty absorption that normally goes the other direction up to the leadership where there’s incomplete information that people either try to fill in or they just act upon what they know. And what is then the expected outcome by the decision makers wherever that is going to just be disappointing and is not going to be achievable. And when the later down the line it’s made visible to clean it up and to get it back on track waste time and it’s difficult for people to get reoriented because of the policies and the cultures and forth that they’re living in have been so entrained in them. it everything just becomes more complicated.

SEGMENT 3: THE TELEPHONE GAME WITH A MACHINE IN THE CHAIN

Kevin Novak: So in the I wanna build off of and this is really a second point that you raised to me when we talked and I opened the episode with an example of that and I told you called it the telephone game, right, with a machine in that chain. In Organizations, published in 1958, March and Simon described uncertainty absorption. As information moves towards the decision maker, the inferences replace the actual evidence. So, playing the game of telephone, each person in the chain passes along their conclusions rather than the raw material. The decision maker receives something clean, confident, and stripped of its limitations. So you describe leaders who assume they know how information came to be and what its limits are. And leaders who are embarrassed to reveal what they don’t know, which I know you and I see far too often. And who, because of that embarrassment, right, never manage the interaction at all. They just accept the answers. So let me give you a version I saw a few months ago. A senior leader at a client organization ran months of grounded strategy work through a language model. Real work, real interviews, real data. And the output was wonderfully polished and it went up to the leadership team as a recommendation without much vetting. And this leader hadn’t anticipated a single one of the questions that followed. Which just when I talked about this hurt yeah, I was befuddled as to why one wouldn’t have taken that pause. And the tool had absorbed all the uncertainty on the way to its own conclusion. And nobody asked it to get give any back. So my question here is in an organization where AI now generates a growing share of the information, the knowledge being absorbed, does that telephone game get better or worse?

Adrian Wolfberg: Yeah, well it gets worse. I don’t think this is a controversial thing that people have talked about and have written about the lack of transparency in being able to figure out things that people have should be asking and they have asked, well before even AI became a thing. And that is and this is something certainly I learned when as a naval officer, it’s okay, well, where did you get this information from? You know. How did you interpret it? what assumptions did you make? and excuse me, what excuse me, what assumptions did you make between the original information that you got and what you’re telling me now? And then what do I need to either reinforce what you’re saying to act on it or what do I need to refute? what new information do you need to do need to refute so that I wouldn’t act on it. So I mean those sorts of questions you know have always been there which people should always been asking even before AI but these become you know much more important now. And yeah so that it’s it’s a shift from you know evaluating just the answer. Or, what I mean by that is if you say something and someone may not understand a term, or they may not understand the flow of the sentence, the kind of traditional way is historical approach has just been, well, can you tell me what you mean by that word? Or can you tell me what the sentence can you say the sentence in a simpler way? Well, it’s gotta shift from that evaluating just that delivery statement to evaluating the path that produced the answer. And that I mean that is especially so much important with AI, but with AI. The I think now the there’s in this age that we’re living in now, there will never be any never will be never there will never be anything that’s purely a hundred percent AI. There will always be some human component of the interaction between AI knowledge output and human it could be a little it could be a lot this is kind of a spectrum of how it operates. So if the knowledge of how it came about is unclear, then either it needs to be cleared up. And if the source and the reasoning and the kind of confidence you know can’t be found then they need to kind of take time we need to stop time and dig a little bit deeper and figure out as best we can what’s going on so that you have enough information about how the path of this knowledge to trust it to make a decision. So this path of knowledge creation, that’s been something that, I’ve been interested in the last forty years and certainly in the last twenty years I’ve been researching this. And yeah.

Kevin Novak: So Adrian, I wanna it’s kind of a digression, but you hit on these points. So obviously looking back and knowing that you’ve researched this as well on the industrial age mindset. Right. And in many ways. Many people were viewed as cogs in the wheel on the production line and weren’t looked at for their thinking or the information that they may have. And leadership teams would feel overconfident that they had all the information to be able to make business decisions or even decisions from that floor. And groups of people in a room, to your point, on whether someone asks a question or uses a technical term and then there’s not clear understanding, there’s a cascade element that if somebody does seemingly understand, that the others will acquiesce and agree because they’re they have trust or they have respect for the person who seemingly understood, whether they did or not, there’s some level of confidence that they apply. And there’s a body of research on this from people who study pilots and control rooms. And in 1997, Raja Parasuraman and Victor Riley published a paper in Human Factors called Humans and Automation. And they found that the problem is rarely that people refuse to use automation. The problem is misuse. And I have a complex question coming with this. So people come to rely on the system past the point where it deserves it. They stop checking, they accept the output because the system has been right before. And the researchers called it automation bias. And it shows up most exactly in the people who are busiest and most confident. And there are two things that I want to hit on with you. What are the influences of that today? But then also the human factor dynamic of it where in those power positioning group settings, people will acquiesce and they will stay quiet, opposed to actually having that information boil up and actually be part of the conversation to the point of getting to the right solution. So what should they actually do different in those kind of moments?

Adrian Wolfberg: Well to the second situation, this is gonna be influenced by the culture, the norms of the organization and the individual who’s in charge or who’s got a position of authority. And it’s then incumbent upon the individuals in the room to assess whether there’s that safety to be able to say something. And not only so there’s a couple of issues. There’s you know the psychological safety of the people of the leaders or the organization who are willing and have an openness to having people share their views. Without some kind of retribution. So that’s certainly one thing. And then there’s another thing from the, again, both from the psychological and the this the social, the cultural aspect of the organization, about how leadership is viewed. And if leadership is viewed by either the individual or the culture as the person who is right, then even if there is psychological safety in raising voices, there you know would be an acknowledgement that, well, I did my best to raise a point. I know that this regime just doesn’t listen if it’s already not part of the decision making approach, but I did my best and forth. So there’s those both kind of you know there’s both of those things go on and there’s plenty of examples where even in different cultures, different societies where there’s systematic biases, that are entrenched in that perception of leaders and need leaders being the ones who know and they’re not and they should not be subject to being challenged. you know we’ve seen that the outcome of that in airline crashes even in the military or even in business or in medicine. it doesn’t it’s not it’s sector insensitive, it’s just a function of this kind of gets to this issue of confidence. And the way I want to introduce that is that there’s a th a human need for people to feel that those who are leading them are behaving in a way that they can feel competent in and not worry about it. And the value of having confidence in those who are pulling the levers to get things done is that you don’t have to worry about the kind of incorrectness or the chaos or whatever, and you can focus a hundred percent on what your job is. And if you don’t have confidence, then that ratio of worry to performance gets changed from a hundred percent performance, zero percent worry, to something much less, it could be just the reverse, or that would be the worst case situation. And so everybody, I mean this is well known and that people who are at the top and whatever level the or of the organization or board of directors or whatever, know that there’s a psychological and sociological advantage for individuals to extol confidence. that is confidence creates progress, it creates movement, because it provides people with the motivation to focus 100% on the mission and not you know worry about many things like their job or so forth. The problem is, and there’s so there’s nothing wrong with confidence, but I’m gonna caveat that. The problem is when the confidence itself is empty, and I call that apparent confidence as opposed to earned confidence. So, when I flew in jet aircraft in the Navy, I was the mission commander of a seven-member air crew. We each had very specific responsibilities. Mine happened to be about the overall mission. I wasn’t the pilot, but I was in charge of the pilot and everybody else. Yet the pilot could kill us all by crashing the airplane and not all sorts of things. So the crew had to feel confident in the pilot’s ability so that they could focus a hundred percent on their tasks that they’ve been trained and they have experienced on. and if and that could only be earned by experience or by reputation of this pilot. Demonstrating that they actually knew what they were doing, especially in emergencies and stuff like that. But when that’s not the case, when the confidence is not earned, because there’s just an impression of people being confident, and yet they don’t understand what’s really going on. like in my case, I was the mission commander, so I had to understand everybody’s job. Everybody’s capability and limitation in order to synthesize everything and put it together to achieve an objective. If that’s not the case, for a variety of reasons, which cuts across mission issues, organizational issues, social issues, individual psychological capabilities and forth, emotional aspects, then that confidence is not is just apparent. And it could lead to disastrous results, even though it’s giving people the impression that they can feel confident that the person knows what they’re doing, but yet they really don’t. so that’s just another issue

Kevin Novak: Well, and Adrian.

Adrian Wolfberg: That comes into play. And yeah, go ahead.

Kevin Novak: Right. Well, because it’s just I want to bring this up in that you hit on psychological safety. I love Amy Edmondson. Think that is such a critical element in organizations. And it’s not just for good leadership, it’s overall health of the organization. And to that is trust, right? That psychological safety creates the trust, and to your point, the confidence level. Is not just perception. When that trust is there is that relationship that yes, this person knows what they’re doing. I trust them that they have the information they need, and therefore, to your point, I’m gonna follow them. And it’s an interesting element for me in that so many organizations, so many leaders struggle with trying to create that type of environment. And often the people who are working, to your point, there’s a perception that they know what they’re doing. But there are some life and death decisions that person is making and they don’t have the full trust, but they’ve been told to trust that person that they know what they’re doing. How do you what do you think about that? And why is that such a gap in a society that has grown overly complex?

Adrian Wolfberg: It’s a shift of at the leadership level, what has historically been knowledge and expertise at the top, which is what has been the antecedent to a perception of confidence, there’s a shift now where I’m arguing that it’s the leaders, that’s not really important anymore for a number of reasons. And what’s mostly important is setting the conditions so that time can be managed better, so on the front end, better understanding, a fuller understanding of what the issue is can be attained, that those on the edge of the organization who are more likely to detect and observe. Changes that occur in the environment and in the reality and how that would affect a decision that was made at time zero back in the past by the leader to be able to you know you rethink and adapt and modify what’s going on and why that’s important, you’d think that would be important anyway, even before AI. And it was important then, not that it’s was that successful in being achieved, but why it’s important now in the era of AI is because of the speed in which information can be processed. and the things that humans have which AI doesn’t have are gonna be needed more now than ever, especially in the face of this time pressure. So, what are those things? This what goes along with this front-end time management, what is really that all about? That’s all about framing. That’s about figuring out of the universe of what’s happening in the world, what things are important of that universe of things. And the a lot of that has nothing to do with data. That’s been captured and put in databases or put in whatever social media or whatever, but it’s knowledge about how people think, what competitors are thinking, what plans are being thought of, what constraints are operating in the world, what barriers are could affect this and this. These are things that only humans have an understanding of 100% of the context that’s relevant to them. That framing is what needs is the activity that occurs, the human activity that occupies what goes on in the front end where you take time to do that. And that framing has component parts to it. part of it is understanding what the problem is, and that goes to being able to dissect or decompose the complexity of problems that need to be thought about. And then to be able to figure out what elements of the problem are best matched with the kind of amount of AI. so you kind of have like kind of a Rubik’s cube in a sense of the right mix and match of human and AI contribution, which is going to be different depending on the nature of the problem. Simple versus complex problems are going to vary in what the proportion is. So that kind of drills down what that framing is. And then there’s this idea that everything is changing so quickly that a decision that’s made at time zero has to be reevaluated or has to be observed and has to be detected over time. So this is part of the leaders setting the conditions so that the organization has the detectors, the observation of what’s going on and can make the inputs into how these what these changes are occurring and then to if necessary modify the mixture or the ratio of AI and human intelligence to address that particular issue. Embedded within that is critical and creative thinking to be able to think about these sort of things that are strictly kind of human elements in the environment. Another you know human thing that needs to be thought about on the back end is empathy and that is what’s now discussed that people are openly talking about now is not the empathy that I’m going to talk about. What people talk about now is how AI is getting so good that the person who’s sitting in front of the machine, like I’m sitting in front of my laptop right now. And if I was to go into a large language model and have communication over time, there is the realization that the large language model would present itself with a certain degree of empathy in my context. If I was revealing myself in that way and wanted it. I’m not talking about that. Yeah, that’s a problem where presumably AI literacy would improve. But I’m talking about something on the back end where the leader and those people involved with processing changes in reality and coming up with solution approaches. Come up with a finally a decision to think about the implementation of that decision, which that we should always have been doing, even before the era of AI, to think about how that’s gonna impact people who are gonna be subject to that decision. Well in the area in the era of AI, this becomes accentuated to the nth degree because depending on how much the ratio of AI contribution, AI is not gonna understand the empathy of large bodies of people who it’s never been interacting with. And the lead this is incumbent on the leader then to have that competency of empathy of the decision that they’re making. again, it’s something that should have always been there, but it’s now gonna become more important. Accountability is another thing, and that is again, that’s not something that’s new. accountability has always been there.

Kevin Novak: No, but it gets left on the floor all the time.

Adrian Wolfberg: And right. And this will become much more visible in the era of AI if it is left on the floor. and this the you know the downside is you can’t really scold or whatever AI. you can say whatever you want to, a large language model or whatever model that’s being used, but that’s not going to fix the accountability of it’ll be the human who’s going to be accountable. So accountability becomes much more important in terms of what consequences occur. And all these things that I’m talking about, they’ve all been there in as part of human civilization before. And it’s gonna take a while for the importance or the increased importance of these to make their way through the expectation of how people develop and how they should behave and forth, because we’re just we’re in the middle of something new right now.

CLOSING (PART 1)

Kevin Novak: We’re gonna stop there for today because what comes next deserves its own episode. Everything that Adrian has shared so far sits on the front end of a decision. The information that gets laughed off in a Pentagon briefing room and while still reaches the Chairman. The auditorium, where a leader says, You own this problem, and I authorize you to solve it. The line between uncertainty, which more information always can fix, and ambiguity, which only more time and perspectives can. And while the telephone game that got a new player who never gets tired. Next week, part two goes to the heart of his book: how a leader understands a problem before touching the tool. What judgment looks like when it is measured and practiced. Why the leader has to become a shepherd of critical and creative thinking. And the word his next book gives to something every organization feels, and while none of them can name: Drift. If you found today’s episode valuable, subscribe to the Human Factor Podcast wherever you watch or listen. So part two shows up next Thursday without you having to look for it. Leave a rating and a comment and share this episode with your leadership team, especially with whoever in your organization is receiving the answers. Subscribe to my Ideas and Innovations newsletter at 2040digital.com or on Substack for weekly frameworks and research on why change succeeds or fails. Connect with me on LinkedIn, where I post regularly about the psychology of transformation. Until next time, remember, it is never the technology, the market conditions, or the strategy that makes change and transformation successful. It’s the people. And the more deeply you understand the human factor, the more likely your transformation is to succeed. This is the Human Factor Podcast. I’m Kevin Novak. Thank you for watching or listening.