Can AI Solve the Problems That Really Matter?

AI is getting faster, smarter and more capable.

But are we asking it to solve the right problems?

In this episode of Leading People, Gerry Murray welcomes Chris Soderquist back to explore what he calls Systemic Intelligence and why it may matter more than ever in an AI-driven world.

Chris argues that many of the challenges leaders face are not really standalone problems at all. They sit inside wider systems, shaped by relationships, feedback loops, assumptions and mental models that are often invisible until we learn how to see them. 

Together, Gerry and Chris explore why organisations can end up having to solve the same problems over and over again, why intelligent people can become trapped by the models that once made them successful, and what happens when powerful technology is applied without enough understanding of the system around it.

Along the way, they use examples from everyday life, business, public health and AI to make systems thinking practical and accessible.

If you lead people, make decisions, work with AI, or simply want to improve the quality of your thinking in a more complex world, this episode will give you plenty to reflect on.

Connect with Chris on LinkedIn

Visit his website

Enjoyed this episode? The Leading People Community is where these conversations continue to flourish. 

Free to join — Book reviews, Leading People Café recordings, Podcast episode previews, and a growing group of active HR and L&D professionals, leaders, and coaches who take their work seriously. 

Join the Leading People Community

Connect with Gerry on LinkedIn

Transcript

Copy Transcript Copied

Welcome And The AI Question

SPEAKER_01 0:00

Welcome to Leading People with me, Jerry Murray. This is the podcast for HR leaders, LD professionals, and senior leaders who want to bring out the best in themselves and the people around them. Every other week I sit down with leading authors, researchers, and practitioners for deep, honest conversations about what great leadership actually looks like in practice. And how your own mindset, behavior, and presence shape everything around you. If you want thinking that challenges you, tools you can use, and conversations worth returning to, you're in the right place.

SPEAKER_00 0:49

So I think AI is here. The question isn't is it, you know, do is it good or is it bad? It's how do we use it in a way that can actually accelerate and improve our ability to learn. And this is back to adaptive challenges. Most of the issues that we're facing today as a species, climate change, um, rising authoritarianism globally, polarization, um, you know, planetary boundaries being exceeded, most all of those that are really important and crucial for society are adaptive challenges. And adaptive challenges require systemic intelligence.

Meet Chris And Systemic Intelligence

SPEAKER_01 1:29

My guest today is Chris Soderquist. Chris actually joined me way back in 2021, in the early days of this podcast, to talk about systemic intelligence. And a lot has changed since then. AI has moved from something many of us weren't even aware of then to something increasingly embedded in how we work, how we learn, and how we make decisions. But Chris argues that this makes our own ability to think systemically more important, not less. So in this conversation we go right back to basics. What does systems thinking actually mean? And what is what Chris calls systemic intelligence? And why do intelligent people and organizations repeatedly end up solving the same problems over and over and over again? And what's the difference between a routine problem and an adaptive challenge? And perhaps most importantly, where does AI help us and where might relying on it too heavily actually weaken the thinking we still need to do ourselves? Chris brings some wonderfully practical examples along the way from coffee addiction and the movie Groundhog Day to COVID vaccines, malnutrition in Peru, and World War II bombers. So let's get into my conversation with Chris. Chris

From Groupthink To Systems Work

SPEAKER_01 3:06

Sotoquist, welcome back to leading people.

SPEAKER_00 3:10

Thank you, Jerry. It's a pleasure to be back.

SPEAKER_01 3:13

So the world has changed quite a bit since we last spoke in 2021. And you've recently published some excellent papers on systems thinking and what you call systemic intelligence. And you've also highlighted some of the benefits and challenges of using AI, particularly with that systems thinking frame around it. But before we get into that, and that's what we're going to really get into today, but before we get into that, please tell my listeners a bit about yourself and how you got to being an expert on systems and systems thinking and developing what you call sysik.

SPEAKER_00 3:50

Okay. Uh thanks, Jerry. So um without boring too much here, I think it's that the history of this makes some sense. So um when I was in college, um, I was getting a combined degree in um mathematics, applied mathematics, and social systems. And as particularly, I was interested in international relations. When I found out and discovered this concept of groupthink and how we, you know, had a group of uh individuals in the government here who could be thought of as some of the smartest in the world, uh, would actually create the Bay of Pigs fiasco, which we have here. And then that same group of people could get together and make decisions around the Cuban Missile Crisis and solve it. What I really wanted to figure out was how can we take a group of smart people, put them together, and create decisions that are actually positive, beneficial, and systemic in nature. And so ever since then, I've been kind of in pursuit of doing that. So I've been a statistician. I worked for the um Bureau of Engraving and Printing here in the United States, which is the um government manufacturing arm. It's an interesting institution because it was both manufacturing and government organization. And that place um produced money. Uh they didn't let me keep any of it, so that was one of the reasons I left. But um what they do is um, you know, a process that's a manufacturing line. So as a statistician, I was trying to solve problems there, and I was pointing out that the defects that were occurring in a system was coming from some upstream problem that they needed to solve. And instead, they wanted to just focus on how to remove the defects with the end of the system. And I realized that they weren't making decisions based on data and information, and instead, they were making decisions based on the way they'd always done them. And so that led me to think that organizational development and working on the culture was what was needed. And that was a double loop shift, which we'll talk about a little bit more uh later on. Here was my mental model of what was going on in that organization was inconsistent without the way the organization was actually working. So um I then tracked down um an organization uh called High Performance Systems, which had Barry Richmond in it. And it was an organization that um focused on helping people apply systems thinking to complex issues. And Barry was an instructor at Dartmouth College, and he'd worked with a lot of other individuals in the uh system thinking space, um, including uh Danella Meadows, who did the limits to growth um work in the future in the past. He studied under Jay Forrester, who actually came up with the concept of system dynamics. So I worked with him for many years until his untimely death. And since then I've been an independent consultant. And I've worked with um organizations ranging from Hewlett-Packard, Northwestern Mutual, um, you know, Fortune 500, Fortune 100 companies. Um, I've worked with um the CDC here on violence prevention, and I've worked with the uh country of Somalia on trying to reorganize its system, moving away from a centralized government to a decentralized, federalized system. And so in all of those areas, what they're working on are systemic issues. They are interconnected, interdependent problems that um have people in that system who see things differently and need to start working together in a better way to create the results they're looking for. And so I've been using a range of tools to help people to do that over the course of my career, ranging from group facilitation, um, getting people to develop the skills to have better conversations, which one of your guests, Craig Weber, talked a lot about conversational capacity. Um, and I've been working using um simulation models and simulation tools to do that. And that's one of the uh ways I think that people need to um develop better systemic intelligence is to apply and work through simulation models and to think about how the systems are actually fit together. So I don't know if that was uh a useful background or not. That's the that's kind of how I got here.

SPEAKER_01 8:05

We'll be right back after this short break. If

Subscribe Share And Join Community

SPEAKER_01 8:12

you're finding value in this conversation, please subscribe wherever you listen to your podcasts, and perhaps share this episode with a colleague who'd find it useful. Word of mouth is how good podcasts grow. And if you'd like to be part of the conversation beyond the podcast, I'd love to have you in the leading people community, a growing group of HR leaders, LD professionals, and experienced leaders who gather regularly for live sessions, discussions, and the kind of thinking you've been hearing today. You'll find the details in the show notes. Everybody's

Systems Thinking 101 Without Jargon

SPEAKER_01 8:51

story is is interesting, uh Chris. Over the years I've come to appreciate a lot of how these stories evolve from uh something that challenged a person and then they went to explore something and then they found it interesting, and it it created a whole career and opened up lots of uh possibilities for them. And you talk about these simulations, and I suppose if I think about that, and I also think about a point you made in one of your articles, which we'll make available for everybody if they want them afterwards. But one of the points you made about people misunderstanding what systems thinking is, and I think the one of the points you make is just because you work in what's called a system or you talk about IT systems or whatever, doesn't mean you're a systems thinker. And I was hop hoping that maybe what we'll do is spend a few minutes for anybody out, this is really important stuff, that's why I want to do this. Uh, for anybody who's either leading other people, grappling with how to integrate AI into workflows, or or simply just young people out there who are curious about you know how they're going to be able to make a difference and impact. So I was wondering, could we kind of start with a little bit of systems thinking 101? Uh you know, let's take the jargon out because there are people like yourself who can build these very advanced simulations with the models, but actually the core basis of systems thinking is not that complicated to understand if you think about it from some of our everyday examples, whether it's our body or whatever. Maybe you could just unpack some of those principles for my listeners.

SPEAKER_00 10:35

Yes. So um I have been in this field and it it's been called systems thinking for a long, long time. And I've actually started calling it systemic intelligence because really it's a capacity for applying a systemic approach to problem solving. Um, so just because you work in the education system or you work in some processing system doesn't mean you're actually applying systems thinking. Systems thinking to me is systemic intelligence is the capacity to understand how the structure of a set of interconnected systems or systems of systems or even the structure of a human being is generating the results that they're looking for, they know that we're trying to improve. And so, for example, let's say that um my daughter's grades in math are poor. I might have a mental model that the way to fix that is to incentivize her. I'll give her a uh, you know, uh, if she's young, you know, some ice cream or something like that, if she gets an A on a test. And that helps for a while. But um after some time, I find that an ice cream cone isn't enough. Now I need to like take her out for pizza with some friends. And then before that, I've got to give her money. And what it turns out is that my mental model of what's generating performance is inadequate to the task. And so I can think about okay, let's double loop this. And double loop really is kind of checking about the mental models that we have. Let me think about this again. And I could come up with a better mental model that what we really need to do is to develop intrinsic motivation and so figure out ways that um she sees math relating to something that's important to her, maybe some hobby she has, or maybe like some career that she wants to get into. And all of a sudden, now she's motivated to do it without the rewards. So rather than getting her hooked on rewards, I'm actually getting her to solve the problem in a different way and actually getting to improve her performance. That's a mental model, and that's taking a systemic approach to it, really thinking what are the underlying deep-rooted causal structures that's generated performance. Um, another example is polarization. So we're trying to figure out how to work on polarization. Typical mental models that people have are sort of linear and factors thinking, like here's all these factors. It's like, you know, grievances, it's like the information bubbles, it's like all these other things, right? And it turns out that if you think about the structure of what actually generates it, it's really, you know, conversations between people that actually can reduce polarization. And um in France, I think there was a study recently where they looked at to box, and that's I'm probably pronouncing it in per imperfectly there, but tobacques are like these stand-up coffee shops. And to box, you know, people would come in in the morning before going to work and they would talk. And then they would come in at lunch and they would also come in and talk. And then you even go there after work and chat.

SPEAKER_01 13:49

The cigarette shop. They were for selling cigarettes, basically.

SPEAKER_00 13:52

Oh, okay, for selling cigarette cigarettes. Okay, so they'd be standing around. And um, you know, what would happen is if um a community, an area actually shut, you know, one of these coffee shops or these cigarette shops um shut down, um, what would happen is they would see an increase in right-wing extremism immediately after. And so you could think about third spaces and the amount of conversations per third space as generating or reducing polarization. And then the levers become very different. It's like, okay, what third spaces are out there? How do we um you know bolster them, keep them in business, add some more, think about making sure that what they're doing is generating more um, you know, positive conversations that can reduce polarization. Again, that's an operational picture. So to kind of summarize, our mental models are what we use to make decisions. And this goes back to why I was interested in this whole thing in the first place. And our mental models typically are anemic and inconsistent or incongruent with the way the real world works. And so systemic intelligence is the capacity to develop mental models that are congruent with the way the world works so that we can make the world work the way we'd like it to work better.

SPEAKER_01 15:14

So I was uh as you were talking about your daughter and you started talking about um having not ice cream not working and then having to take her for pizza and other things, I I I was tracking that she was get becoming a teenager at that stage, you know, and then you were out there hanging out with her friends and that and all that goes with it. I would I I just wanted to get to a couple of things before we unpack some of that. Um in within a when within a system, a system is characterized by a few factors like relationships. You talked about relationships already, and there are feedback loops always in the system. Sometimes they're amplifying or reinforcing, sometimes they have the opposite effect. And then you have this notion of time, you know, um, the time delay between cause and effect that people don't always understand. And then you have notions of stocks and flows. Uh what what does that mean for for for the the average person out there to understand a system? Because you know, before they start to to go into the ment beyond the ment into the mental model, they they it's worth knowing that that these these things are all around us, you know. Um there's classic examples from your thermostat to whatever and how a system functions. And once you understand that, then you can start to question the mental models that hold that in place. Is that you know what is the structure will drive behavior in a system and these these factors? So maybe just can you can you elaborate a bit more on those very basic principles so that anybody listening now can really start to think, oh, that's a oh, there's a system, oh that's a system, there's a system at work, there's a system in my family.

SPEAKER_00 16:56

Yes. So

Archetypes Feedback Loops And Coffee

SPEAKER_00 16:58

um one of the um contributions, I think, to the world from the field of systems thinking has been um the concept of a systems archetype. And a systems archetype is a type of systemic performance, behavior pattern that um repeats itself at a variety of different um levels of society and across it. For example, I'm sipping coffee now. And coffee, um, I drink it because my energy level goes down. So I drink coffee, my energy level goes back up. That's a nice balancing loop. Um, I'm trying to keep in balance or counteract the low energy level by drinking coffee. But what often happens is that then the energy level goes back down, and I have to drink more coffee. And um I become addicted to coffee, which actually keeps me from doing other things by changing my lifestyle, sleeping better, taking more breaks during the day, just basically a much better well-being. And that would be the loop that I should get into, right? Is I should be like working on better energy management throughout the day, not like getting hooked on coffee, but I get hooked on coffee. And that creates a vicious cycle of having to drink more and more coffee. So I end up looping like Bill Murray and Groundhog Bay, where he's actually drinking the whole pot. Um that's called shifting the burden to the intervener. Well, that happens at the individual level, but it also happens in organizations. Organizations figure out that they're unable to solve the problems that they're working on. So they hire external consultants and consultants come in, give them the answer, and then they try to fix something based upon what the consultant says. Next time a problem emerges, they have to hire a consultant again because they haven't built the underlying capacity to actually solve the problem themselves. They haven't built their internal capacity. And this even goes up to the national level, right? So working in a lot of countries now, and you know, USAID shut down here. Um, and again, I don't want to get into the politics of it, but it's it's horrible. Umganizations or nations around the world have gotten to the point where they were expecting and needing this external source of funding. And with that goes away, they don't have the internal capacity to solve the problems that they would need to solve. And so again, they've shifted the burden to the external aid rather than building their own capacity. So this happens at all levels, and that's just one example of it. But in all of those cases that you take a systemic approach, if you start thinking about the feedback loops, the time delays, when some of the feedback loops become more dominant and you get stuck in the hamster wheel, continuing to have the same problem come up, that's a way of having a systemic approach or systemic intelligence of problem solving.

SPEAKER_01 19:48

Okay. So

Single Loop Versus Double Loop Learning

SPEAKER_01 19:49

we're really getting in now to another question I wanted to ask you about, which is uh what is single loop versus double loop learning? And uh eventually I think we can segue into the distinction between a routine problem and an adaptive challenge uh if we can sort of segue those two together. So you you were given the example there about the coffee. I guess that's your single loop solution. Uh low energy levels. Let's just take that one then and run with it a bit. And then let's get into the organizational uh or even national, international challenges around routine problems and adaptive challenges, and then we'll then we'll get into where AI sits and all of this. So let's I'll let you flow.

SPEAKER_00 20:36

Okay. Let me flow. All right. So um the concept of double loop learning um came out of the work of uh Chris Ardris and Don Schoun at uh uh Harvard and MIT. And they were looking at how is it that organizations, again, continue to see the same problems arise. And what was occurring was that they were using the same mental model to solve problems. So if you go back to the coffee example, right? So here's the problem. I have low energy and I need to fix it. So my mental model is very, very short term and it is very, very um sort of in the narrow space and time bounds. And that's the challenge with our mental models. If I assume that's the problem, then I, you know, the solution is drinking more coffee. I just drink more coffee. That would be called single loop. And you kind of know you're in single loop thinking or single loop mental models when the same problem keeps coming back. You thought you fixed it and it came back. Or sometimes it's like whack-a-mole management, you know, where the mole pops up and you whack it with a hammer and then it pops up somewhere else and you whap it. And um, if something like that keeps occurring, then you probably need to stop and say, my mental model about how things work here probably is incorrect. And a double loop shift is usually expanding the horizons over which you're thinking about the problem. And you think about it over a longer term period, you'll say, okay, I need to create a lifestyle that actually keeps my energy in balance. And that is usually a longer-term fix, takes a while to work on. But again, it's a root cause of the issue that I actually was trying to solve with a single loop approach. And we have that, again, this single loop approach to just about everything. So I worked with an organization, a financial institution, and they were processing, it was they were processing insurance or disability claims. And what was happening is it was getting stressful on the line. And so people were quitting. And whenever people quit, they didn't have obviously enough head count or head staff to meet the demand for processing claims. So they would hire more people. And they found themselves finding that more and more were quitting, so they were hiring more and more and more people. That was a single loop problem because what was actually happening was when they're hiring new people and on the line, that actually took away time and energy from the senior leadership in that group to help mentor these folks. And while they were mentoring, the overall productivity went down. And that actually caused more stress on the line and more people quit. So the solution was really managing stress and working on emotional intelligence and working on mentoring capacity so they could do a better job of keeping the people that were already hired and managing the stress on the line. The singular loop uh solution is always just hire more. And you'll find that in all most organizations. We don't have enough people, so it's hire more. The real question is why are they leaving? How do you solve that problem? And how do you then build the skills and productivity of the current group? And that's more of a double loop shift.

SPEAKER_01 24:00

Okay. And um let's so you have the groundhog example actually coming out very strongly there. And I think you made a point in one of your papers um that the the whole philosophical purpose of that movie was to actually demonstrate that um you keep doing the same thing over and over again, you're going to keep getting this this reaction. And eventually Bill Murray, the actor, gets the message, right? Yes.

SPEAKER_00 24:31

Yeah, that is that's yeah, that's actually a really good movie, and I I recommend it highly. Um, you know, when I do some training, you know, when I do sort of like week-long training, we actually watch the movie. And in watching the movie, you see that he starts off very short-term, very centered, self-centered, and trying to get what he wants. And he, you know, each day he learns more and more about how to get what he wants. But ultimately, he learns that it's not really getting what he wants that's important, but it's actually being the kind of person who does the good stuff that actually gets you happier. And so he gets stuck on this wheel of thinking happiness is driven by stuff that he gets. And he learns that happiness is actually generated by being a person who actually reaches out, helps people, and takes a whole different approach to life. And it is single loop learning all the all the way along. And then eventually he has that double loop shift towards the end of the movie that's you know fascinating to watch.

SPEAKER_01 25:31

Yeah. And the rest of us are are painfully wishing he'd get it sooner. Because it keeps going. Come on, get the message.

SPEAKER_00 25:40

Yeah, you know, he's trying to get this girl and he keeps getting slapped by her because like he he doesn't he doesn't get it.

SPEAKER_01 25:46

Yes, he doesn't get it. So watch the movie, ladies and gentlemen. Uh, you'll you'll get some learning out of it, particularly after you've listened to this conversation and maybe even read uh Chris's papers. Okay, so let's let's get into this uh this concept of routine or technical problems, depending on who's talking about it, and adaptive challenges. And now

Routine Problems Versus Adaptive Challenges

SPEAKER_01 26:11

we get into a little bit the world of Ron Heifetz and also our mutual friend Craig, Craig Weber, he he added something to this. Craig was uh a guest on the podcast in those early days as well. Um, let's expand that now and let's get some examples out there so everybody can go, oh, so that's a routine problem, uh, but we're thinking that's an adaptive challenge, but we're thinking of it as a routine problem and the adaptive challenge, that's why we can't solve the adaptive challenge, it said. So maybe just expand that for us and the implications of not understanding the difference between these.

SPEAKER_00 26:48

Okay. Um, so I'll start off by just describing that systems tend to get stuck in producing the results that they're they're producing now. So people would say, oh, the system isn't isn't generating what we want. Well, the system is generating the actual results that it's designed to produce. So you might think of a system getting stuck in a lower performance orbital. And let's say that there's a higher performance orbital that you want to get to. What gets you stuck in the current um system, the way it's performing, is the um is basically the resources that you're allocating, it's the decisions you're making, it's the relationships and how people in the system are working together. And underneath that, again, are mental models. To get up to the new level requires a completely double-looped shift. And usually what happens is if you're not getting the results you want, you end up solving these routine problems. So if you think about problems in terms of two types, and one type Ron Heif is referred to as technical problems, and as you said, Craig Weber refers to them as routine problems. And these are problems that we love to solve because they're easily diagnosed, what they are. They're um there's usually an expert that knows how to solve them, and there is a routine for solving them, and that's why Craig Weber refers to them as routine problems. Take, for example, COVID. So COVID comes along, and one of the problems that's a routine problem in there, actually, and they might not have thought this at the time, but it really is um creating a vaccine. And the vaccine required actually some AI tools were used to develop the new mRNA vaccines, and they came out really fast and rapid. Well, it's easily diagnosed. We need a vaccine. Um, there's an expert, usually people in the pharmaceutical industry, and there's the scientific method, and they just go to work on it. That was a routine problem. And it didn't happen routinely. We haven't had that particular type of a virus ever, because that was a novel virus. So it required a routine to solve it. On the other hand, there are adaptive challenges. And so if you think of COVID, um, what was an adaptive challenge because it was not easy to diagnose, which was why is it that people aren't taking the vaccine? Um, there was no expert that had a solution to that. And there was no routine to solve it. So we were on totally different territory. So in the United States, we did not have a real uh good uptake of the vaccine. In New Zealand and other places, they did a really good job. They were adaptive. It was an adaptive problem, and some organiz some countries actually did a better job than others. Um, it's estimated, I think, in the United States, if we had had the same level of uptake and in terms of social norms and of getting vaccines as New Zealand, we would have saved half a million lives. So, but that was an adaptive challenge. An adaptive challenge requires something completely different from a routine problem. A routine problem, you know, you know what is the routine, you've solved it, you've got the issue identified, there's a routine, and an expert just get to work and do it. If it's an adaptive challenge, you actually have to accelerate learning. Learning is what you need to do in the system. Some countries did a much better job of learning about how to get social norms and um how to actually get people to take the vaccines than others did. And that was an adaptive challenge. So when you're faced with an adaptive challenge, rather than just get to work on solving it the same old way, you need to solve and actually reframe what the problem is. Now, how does this fit back into that orbital concept? When we're kind of stuck in a lower performance orbital, there's a bunch of routine problems and there's a bunch of adaptive challenges, and you need to solve them to get to the higher level. What we tend to do is we get attracted to solving the routine problems. Let's hire more people. Um, you know, let's like develop the vaccine, let's get to work, let's get dizzy. And we tend to focus our energy and attention on solving those routine problems, and it's the adaptive challenges that we overlook and we don't want to solve.

SPEAKER_01 31:21

Yeah. And and you meant you mentioned this actually, and I think this is where where many of the uh my listeners might uh hook onto this is change management programs are a classic example of how the the the sort of like there's a gravity that pulls people, use this word orbital, it's in your paper, but it's a gravity that pulls the people back down to the orbital they're in in the way that they apply, as you say, um try to solve routine problems when actually they need to get to that higher level orbital and that they need to shift their mental model. Is is that my have I interpreted that correctly, what you were saying in the paper?

SPEAKER_00 31:59

Yeah. And if you think about it, um, you know, we we are capable of doing adaptive learning. It just isn't something that we are sort of gravitate towards. We like to solve the routine problems. We you know, we've got hammers, we want to go around and find nails and start like whacking them with the hand with the hammer. And we don't think about, well, how do we construct something new?

Reframing Addiction Portugal’s Breakthrough

SPEAKER_00 32:21

And that was something that actually in Portugal, there was a nice example of that, I think, in one of my papers, about the fact that um, you know, for years, Portugal had um one of the highest rates of heroin addiction um, you know, per capita of any country. And it had been uh sort of impenetrable, inscrutable. They couldn't solve it. Um, and uh they'd been focusing on um sort of typical approaches, which was make it a crime, um, put people in prison, you know, those kinds of punishment approaches. And they double loop shifted it. Um they did a double loop shift and they thought about like, well, this isn't really a criminal problem. This is a public health issue. When they reframed this issue as a double as a public health issue, they came up with different solutions. Let's create a community, a society with a safety net, harm reduction, other kinds of things. Let's address some of the basic needs of people. And in doing so, they cut the rate of um heroin. I think it was like 100,000 who were addicted. They cut that down to 25,000 in just a few years. So a 75% reduction. And that was basically due to reframing the issue from a criminal problem to a public health issue. That then created the opportunity for learning and doing things differently to get to the higher level of performance. And that's something that's very important. And I want to kind of highlight that is usually when you're stuck, you need to think new first so that you can then build and apply that thinking to get to the new level. And we often don't take the time to say, hey, where are we thinking about this wrong? So we can rethink it, reframe it, and then solve it or resolve it to the level that we need to get to.

SPEAKER_01 34:14

Okay, and let's also unpack the notion of a mental model and just because I one of the things about mental models is that they contain things like beliefs and values and assumptions almost tied to our identity, aren't they, in many ways. And so, you know, and if we think about where we where like things like beliefs and values have very strong energy and locking in forces that that keep us because because our brain likes them, because they've they kind of keep us safe within the world that we've created for ourselves. And it is a world we create, of course, our brain creates our own reality. Um so just talk a little bit more about the mental model uh as a concept, and then what sorts of interventions have you seen work in terms of loosening up belief systems and and challenging assumptions and helping people come to their own conclusions?

SPEAKER_00 35:11

Yes.

Mental Models In Peru Malnutrition

SPEAKER_00 35:13

Um that's a a really good question. And so what um I have found is that people's mental models are, as I mentioned before, very short-term oriented, um, very um sort of narrow in space and bounds of time, boundaries of time, and um really sort of like correlational. A lot of factors are generating it. So, what I do with folks is I help them to kind of map out what is it that's actually generating the performance that they want us um actually perform and improve. So, um for example, when um working on malnutrition in Peru, um there were a lot of different entities and organizations that were focused on that. There was the public health ministry, there was um actually the uh the economic ministry, uh, there were humanitarian organizations that were involved, there were communities, um, the World Bank was involved. And they all had their own sort of individual mental models about how to solve the problem, what was causing it and what could fix it. And so you think about malnutrition, and what was happening in Peru is that um due to malnutrition, um, actual uh adult heights were stunted. So malnutrition at a young age actually stunts growth. And so they wanted to address that. And so the typical mental model is well, we just got to get food to the families. So they would uh humanitarian organizations would get food in and they would give it to the families that had small children, but their mental model was incomplete in that what was actually happening is that the food was going into these families with small children, but the families, because they didn't understand the necessity of giving it to the small children, were actually giving it to the older brothers who actually were earning the income for the families. So it was a mental model that was actually incomplete, incongruent with the way um you know families would actually use the food. So they needed to focus in on some different things. Um that mental model, um, once we figured that out, is they worked more on education. They worked on getting their the parents and the uh the mothers to bring their children to the clinic, they educated them through the clinic, and the actual leverage was actually getting more mothers to the clinic so they could actually educate them so that when they got the food, they would actually solve the problem. And it was this sort of like expansive mental model that was required across the different entities and organizations that were responsible for getting food in to understand how to do that. And so that coordinated mental model or that sort of collective mental model was what was needed to fix it.

SPEAKER_01 38:22

You're listening to leading people with me, Jerry Murray. And my guest today is Chris Soderquist. So far, we've explored what systemic intelligence actually means, why our mental models can keep us trapped in the same patterns, and the important difference between routine problems and adaptive challenges. But this is where the conversation gets particularly timely. Coming up. Chris and I explore where AI fits into all of this. Why is AI so powerful when we're dealing with routine problems? And what happens when we start applying it to adaptive challenges? And could one of the biggest risks actually be that we gradually outsource the very thinking we most need to strengthen? Back to our conversation. Yeah,

AI Helps Routine Work Not Learning

SPEAKER_01 39:18

I I think it I think it does. I mean, I think it's a it's a fantastic example. And I think what uh our my listeners probably are beginning to detect is that you've done a lot of work, not just in ordinary like Boeing and all these other big companies that you've worked with, but you've also done a lot of work in the healthcare space, which of course I believe is something that should interest all of us because uh as my mother always says, your health is your wealth, you know. And if you don't have health, as we know, in and when healthcare gets expensive and that, then that's even worse for people who suffer from poor health. So, but now I want to get to, and I want to kind of put this all into a couple of questions here, but I'm gonna kind of put them into one bucket for you. So, where does AI fit into all of this, particularly when we look at the routine problem and the adaptive challenge? But you also have this great metaphor about putting making sure your ladder is against the right wall. And of course, I guess there are you're either a doomer or a boomer when it comes to AI, they say today. But I think there's there's different ways of looking at how the lack of systems thinking might lead to uh unintended consequences, severe unintended consequences, but there's still time to to deal with it. So let's just unpack where does the AI fit into all this? Because you wrote a brilliant paper on it and um explaining it. So off you go.

SPEAKER_00 40:48

Off I go. Um so um I am not a doomer um or um you know a boomer on AI. I think that AI, first of all, it's here. Um I don't think there's anything that we can do about stopping it. It's not likely to turn the switch off. Um, and one of the things that you understand if you spend enough time in the system thinking space is once things are built, once infrastructure is put in place, that infrastructure wants to stay there. There are incentives in the system to get some return on investment on something that's put in place. So we've been putting a lot of stuff on place, the data centers, et cetera. So I think AI is here. The question isn't, is it, you know, do is it good or is it bad? It's how do we use it in a way that can actually accelerate and improve our ability to learn. And this is back to adaptive challenges. Most of the issues that we're facing today as a species, climate change, um, rising authoritarianism globally, polarization, um, you know, planetary boundaries being exceeded, most all of those that are really important and crucial for society are adaptive challenges. And adaptive challenges require systemic intelligence. They require a systemic intelligence as part of the learning process. It's not the end-all and be-all. Having good conversations like Craig Weber teaches with conversational capacity. You had Frank Barrett on with yes to the mess, having an improvisational um leadership approach. All of those are what we're going to need to solve these complex, interconnected problems that people are referring to as a polycrisis. So that's sort of the big picture frame, is these are adaptive challenges. And what is problematic is that AI is really, really, really good, better than we are in most cases at routine problem solving. So we've got this tool, and I'm back to the hammer analogy, we're going to want to shove it and use it on routine problems. And we're going to basically, and this is my concern, is we're going to atrophy over time because we're spending all this time outsourcing the thinking, actually building the systemic intelligence capacity that we need to solve these adaptive challenges. Um, and AI has a few problems with it. One is um it's mostly, at least now, pattern recognition. I mean, we're looking at trading more sentient approaches to AI, et cetera, in the future. Even then, I'm questioning whether or not it'll have the capability of doing what we talked about about double loop learning. And the reason for that um is because it's trained on the mental models that we currently are using, that's creating the patterns in society that we're currently seeing. And it's reverse engineering the mental models based on the patterns. So if it's focusing in on patterns that have inherently flawed mental models behind it, generating them, all we're going to do is reinforce those flawed uh mental models that have been embedded in the way we're acting and behaving. Organizing as a society or as an institution, as an entity, even as a family unit. All of those things are going to be overlooked and embedded implicitly behind it.

Survivorship Bias And Built In Bias

SPEAKER_00 44:12

So there's a nice example in my paper about, you know, bombers in World War II returning back to Britain. And what they were trying to do is to figure out well, what can we do to make sure that more bombers come back? And so they were looking at the wings on the plane and the fuselage and everything and where their bullet holes were. There was a pattern. And so the typical approach is like, okay, let's just, and this was the initial um, you know, conclusion by the researchers on it was let's reinforce where the holes are. And there was a uh statistician who's named as Wald, um, who actually said that's the wrong place to actually reinforce. That's where you can take a hit and actually come back. It's where there are no holes that you need to actually reinforce. And that is kind of in a nutshell the idea of you can get AI to determine the patterns, the conclusions that it will come up with are often based on our flawed mental models that need to be changed and improved. So I think I think that kind of is one of my concerns is that we've got in AI, we've got flawed mental models that are actually generating the patterns that it's using. And so fairly simple example of this is in writing the paper. And I said I'm no um gloomer or doomer and um you know Luddite when it comes to using AI. I was using AI to help improve the quality of the paper. And I gave it a paragraph about um a consultant that I was talking to in a financial institution who was actually bragging about how awesome it was that they were moving their um different uh computing systems closer and closer together so that the feedback loops and the time delays between decisions and data was uh sped up. But anyway, um when talking about this consultant, I said improved this this um this paragraph. And it came back and it called this person a he. And it was actually a she, it was a woman. And that was just because the data it had been trained on implicitly assumed that if you're a consultant in an organization, you've got to be a man. And that to me is really problematic because we've got all this bias built into the AI systems that come from our flawed mental models. And we've got you know, sort of short-term linear thinking, factors thinking, um, all of the things that systemic intelligence kind of um prevents from being a problem, um, actually is embedded currently in the in the AI. I just find that's gonna be dangerous. And we're just gonna keep solving these routine problems. And if we try to apply to an adaptive challenge, we're gonna create some really bad unintended traffic sequences.

SPEAKER_01 47:04

Yeah, and there's a couple of terms being floating around. Some academics have done great research, and you uh you've added a third term to this this idea of offloading, um, you're thinking, which which typically we did with a calculator, it just helped us do mental arithmetic a bit faster. And then the I can't I can't remember the two guys' name now written paper on this, but um, they also talk about surrendering completely to uh AI, but you've even taken it further because you start talking about total dependence on AI, I think, in one of one part of your paper. So, so how can uh the sys sysIQ help? How can systems intelligence systemic intelligence help us avoid both surrendering and becoming dependent on AI to solve our problems?

SPEAKER_00 47:53

Yeah,

Avoid Offloading Thinking To AI

SPEAKER_00 47:54

um that's a tough one. Um, because I I think that systemic intelligence, again, is one of those things that's developed by um spending time studying, developing a mental model of how a system performs, and then having um that critiqued, either through the way the real world works, through data coming in to modify and improve our mental models, um, or um, you know, computer simulation or mentoring, et cetera. It requires a disciplined approach to developing a capacity that we can have, but it's very challenging to develop. My concern is, and there's a big movement in my field of system dynamics, of creating simulation models and basically saying, here's the problem, let's let AI create the simulation model. And as I said before, that is probably going to embed in it bias of the models that are currently out there, and it may not create a model that actually encompasses a broader, bounded view of things. It may lack some feedback loops that need to be included, it may lack soft variables that need to be included, things like understanding how morale in a society uh in a company contribute to performance, or things like how trust in a society will actually work for or against polarization, things that are harder to measure. And the the approach to that is really sort of uh sort of working on it and trying to understand it and trying to build out that mental model and to test it and improve it. And as they say in my business, you know, all models are wrong, wrong, some are useful. How to actually make it less and less and less wrong. And what I think will happen is we'll try to solve problems and we'll use AI to do it, and we will not go through the you know the muscle building that's required to develop the systemic intelligence. We'll just assume that it's handled by AI. And that to me is really where I think there's uh uh this offloading that's going to occur is we're just gonna offload all the problems to AI. And the final thing I just want to say on this is that um what happens, I think, when you use AI or to any machine learning tools or whatever, is you approach more and more the mean, and you're unable to um sort of expand the boundaries, expand out the way you're thinking. And so the mean is typically less than high quality. And what we're occurring, what's occurring in society right now, we're moving into a world where the standard deviation is going to be much more likely and problematic. We're gonna be running up against things that are problematic and standard deviation, with the every 100-year floods now being like every 10 years, et cetera. Um, so we're going to need to have the ability to move from sort of the average and the mean to thinking about sort of these wider extremes and how to manage and work in those. And so that's that's my concern is that we're going to give AI the problem solving, we're going to lose the ability to think about the extremes, to think about how to expand boundaries. And yeah, then then we're we're we're we're messed up.

SPEAKER_01 51:18

So if if the using the concept of the mean, if the average person out there just tries to solve a problem using uh an AI chat GPT or Cloud or some of these guys, uh it how useful is it to think about that as just a mental model product re you know being presented to you and actually challenging it using some of the principles you've talked about and questioning the mental model that's that the what is being represented in the answer. And have you have you tested this out and does it make any difference?

SPEAKER_00 51:54

Um so a few things. Um, you know, first of all, I think what happens, and this is something we fight in the field of systems thinking and system dynamics anyway, is that people tend to want an answer generator. Whereas any computer simulation model, any model, any mental model is just a set of assumptions that we're putting together to try to understand how a system works. And then they're in our head and we put them into a computer. So a computer really is a mental model. A computer model is really a mental model that's just been quantified, stuck into a computer, simulated. Because sometimes we're just not that good as as um thinking it through. Our mental simulation is really anemic. We could get into that in a bit if you want um in a minute. But um what we really need is um when you're when you're focusing in on um you know the these mental models that are are there, what can happen is you tend to um just expect the computer's results. So you can expect an AI when it comes up with an answer rather than saying, well, that's a mental model that's been generated by the patterns that it's observing, we just say that's the answer, and we'll turn it into a black box. And that will demotivate us to actually question it.

unknown 53:25

Right?

SPEAKER_00 53:25

Because all it's coming out of the computer, it's an AI-generated result, therefore it's unquestionable truth, and we won't question it. And so that's one of the ways I think that we will lose our ability to think more systemically, is we won't even question and say, that's a set of assumptions inside there. Let me try to understand them and challenge those assumptions, which is where the mental muscle comes from for systemic intelligence.

SPEAKER_01 53:53

Right. And I would I think we have to have another conversation about the anemic uh aspect of our thinking because we're a little bit up against the clock

Practical Steps Papers And Learning Labs

SPEAKER_01 54:03

now. But coming to the end, Chris, what are some practical takeaways from my listeners? What are coup some things listeners could do to start to build even just the baseline of systemic intelligence themselves? And I believe you have some special offers. Maybe that's part of the answer because you have uh some interesting stuff that you've been working on, which I think you're prepared to give access to if some people are interested. Is that right?

SPEAKER_00 54:30

Absolutely. So um you can find me on LinkedIn. Um and I also have a YouTube channel, uh Politex Consulting Channel. Um, but you can reach out to me on LinkedIn, and I'm uh willing to share some papers that I have, papers that Jerry has mentioned on this um podcast. Um there's other resources that are out there available that I can point people to. And I'm just now starting up some uh communities, some learning communities around this. One is the Orbital Ship Lab, which will be um looking at how to do systems transformation or guide systems transformation to build those change agent skills. And the other one is the flourishing constellation, which, if people are interested in understanding how these skills are going to help us to navigate the polycrisis, that's another um community. So reach out to me on LinkedIn and um I will uh happily share um articles of papers I have and resources and get you connected to these communities if you'd like to have uh sort of guided mentored practice at developing these skills.

SPEAKER_01 55:33

Yeah, and you have some course material in there, don't you, that uh people can start to follow some guided proper instruction.

SPEAKER_00 55:40

Yeah, there's some material in there that teaches some of these skills. Um, and there's an even there's an interactive exercise called the Rookie Pro Challenge, which is kind of uh a useful one to talk about why our mental models are flawed and how computer simulation can actually improve the quality of our mental models. So there's some interactive free resources and um material you can even use to facilitate groups or teams that you're working with if you want to apply systemic intelligence to the issues.

SPEAKER_01 56:10

Well, I think that's a great, great and generous offer, Chris. And and look, we've covered quite a bit of ground here in this short conversation, and that um we could continue, I'm sure, to explore much further. And at the same time, uh Chris, Chris Sotoquist, on behalf of myself and my listeners, thank you so much for sharing your insights, wisdom, and advice with us here today.

SPEAKER_00 56:32

Thank you.

Final Reflections And Goodbye

SPEAKER_01 56:39

And that's it for this episode of Leading People. If you're new here and enjoyed today's conversation, I'd love it if you subscribed. It's the best way to make sure you never miss an episode. And if you're a regular listener, the best recommendation is always a personal one. So, if this episode made an impact on you, please share it with a friend or colleague. It might just be the best thing they listen to this week. Until next time.