Ep 134: Unpacking the Psychology Behind AI Success with Dr. Gleb

[00:00:09] Hey everyone, it's Bob. Welcome back to Elevate Your AIQ, your go-to source for insightful conversations on human-centric AI readiness, talent transformation, responsible innovation, and the future of work. Today, I'm joined by Dr. Gleb Cyperski, better known as Dr. Gleb, founder of Disaster Avoidance Experts and author of the new book, The Psychology of AI Adoption at Work, From Resistance to Results. Dr. Gleb built his academic and consulting career studying cognitive biases and decision science, and he brings that lens to one of the biggest challenges.

[00:00:39] Dr. Gleb's biggest blind spots in AI adoption, the gap between what leaders assume motivates their people and what actually does. We dig into why organizational change efforts often miss the real problem, how executives and everyday employees experience AI adoption so differently, and what it takes to get people building AI into their own workflows rather than waiting for it to be added to them. It's a super insightful conversation as well as a sneak peek into Dr. Gleb's new book, Out Now, which itself is based on his extensive research into a widespread challenge.

[00:01:09] As always, thanks for listening and thanks for being part of the Elevate Your AIQ community. Let's go talk to Dr. Gleb. Hey, everyone. Welcome back to another episode of Elevate Your AIQ. This is your host, Bob Pulver, and today I am really looking forward to my conversation with Dr. Gleb. How are you today, Dr. Gleb? I'm doing well. Thank you. I appreciate you having me on the podcast. Look forward to talking about AIQ and how it impacts human psychology.

[00:01:36] Awesome. Yeah, likewise. Yeah, you've done some amazing work in this space wearing a number of hats. And so, of course, before we get into all the fun AI stuff, I thought you could give my listeners just a little bit about your background and some of the research you've done and books you've written and, you know, the teaching that you've done. You've really had an impact in a number of disciplines. So I'd love to hear about it.

[00:01:59] Sure. I'm definitely happy to do that. So yeah, so my background is in how to make good decisions. And I came to that when I was turning of age, coming of age, around 18. I was 18 in 1999 when the dot-com boom was booming and there were companies like Webvan and Pets.com that were very popular and praised in the Wall Street Journal. Their leaders were praised in the Wall Street Journal.

[00:02:26] Then, as I turned 21, then in 2001, 2002 was a dot-com bust. And all the companies and leaders that were praised in the Wall Street Journal now began to be criticized in the Wall Street Journal. But the leaders didn't change. What changed was the situation. And so that helped me realize that their decision-making style, we don't really know much about how to make good leadership decisions and how to avoid bad decisions.

[00:02:57] And neither does the Wall Street Journal. So the leaders are praised for their decision-making and criticized for it. So I decided to study this topic. And the relevant area here is cognitive biases. So I decided to study cognitive biases. And cognitive biases explore how do human beings make decisions and make decisions badly, specifically. And then my expertise specifically is in de-biasing, how to make better decisions, addressing those cognitive biases.

[00:03:23] It doesn't have anything to do with racial discrimination or anything like that. That's not the kind of de-biasing that I do. I do the scientific de-biasing, where you look at the science of decision-making and help people make good decisions. So that's my expertise. So I got a PhD at UNC Chapel Hill. Then I taught for seven years as a professor at Ohio State in the Decision Sciences Collaborative. And so that's where my expertise comes from.

[00:03:51] And all this time while I was teaching and training, I was also doing consulting on the side for companies and training companies on making good decisions in the future of work. So specifically, that's my area of expertise. How do you adapt to the future of work? How do you make good decisions in the future of work? What is the science? What is the psychology of figuring out how to adapt to the future of work?

[00:04:14] So I wrote a number of books on this topic, like Never Go With Your Gut, How Pioneering Leaders Make the Best Decisions, and The Blind Spots Between Us. So I'm well known for those books. And so folks might have read those. And then I also appeal regularly in Harvard Business Review, Fortune, Forbes, The Hill, New York Times. The New York Times has called me the office whisperer for my work helping companies adapt to the future of work.

[00:04:41] And so that's another area, other places where folks might have seen my name and heard about me. And so my newest book out is called The Psychology of AI Adoption at Work. It's a peer-reviewed book. It's from Georgetown University Press. And it's out now, available on Amazon, Barnes & Noble, your physical bookstore, probably airport bookstores, wherever fine books are sold, as people say. So that's a little bit of my journey. And I'm happy to go into any aspect of that that you want me to go into in more depth, Bob.

[00:05:11] Oh, yeah. I think we've got a bunch of threads to pull on all of that great work. And well, first of all, congrats on the new book. I'm looking forward to reading it. And I did read your article around the office whisperer. I think the context was you were helping people with their return to work. That was the future of work at that point. At that point, it was the future. And now it's the now. But yeah, you know, I think that's going to be one of the topics that we'll probably get into

[00:05:40] because I feel like as we think about the talent space, the labor market, and we think about where to find the right talent for the right jobs, certainly there can be both advantages and disadvantages to looking for talent that can be remote, that can be global, work from anywhere. And it also may have advantages and disadvantages for, you know,

[00:06:08] the future workforce and the younger generations in the workforce. So I think one of the things, you know, we can dig into when we talk about, you know, AI at work, and I love the work that you're doing for so many reasons, but one of them is because I've been spending a lot of time talking to, well, guests on the show and clients and others around sort of shifting from efficiency to effectiveness.

[00:06:38] And as we lean into these sort of durable human skills and we look at, you know, people are using terms like, you know, taste and judgment and things like that, ultimately the goal seems to be, does this help us make better decisions or not? Where to invest, where to grow, where to divest perhaps, you know, who to hire, who to put in a leadership pipeline, you know, all of these decisions need as much sort of fact-based, you know, data-driven decisions.

[00:07:06] And we need to mitigate the biases wherever those biases exist, whether we pin them to humans and specific humans, like the leaders you talked about, or people claiming that, you know, AI is exhibiting bias, in which case we need to dig underneath that to see where it learned those biases, things like that. But when it comes to, you know, AI adoption, can you just talk a little bit about why ultimately this challenge

[00:07:36] that most organizations are experiencing is fundamentally sort of a psychological, you know, cognitive challenge? Happy to. And so let's talk a little bit about the research. So very research-oriented. There was a recent study out from MIT showing that something like 95% of AI pilots don't show return on investment in terms of the resources invested into the AI adoption. Now, why is that? What's going on here?

[00:08:04] AI is a very mature technology, comparatively speaking. I mean, you have news stories about it being able to hack independently into quite secure websites. So the technology itself is very high class, high grade, well-developed. So what's going on? Why is it not being adopted widely? And so that's the challenge. And what we need to understand is that AI is different from previous technologies.

[00:08:32] When you think about something like adopting a CRM, or an ERP, or HR management technology, or accounting software, the difficulty with that is learning the new tool. It takes a lot of learning to click buttons here rather than there, then learn about how to use this new tool with your existing processes. It's a challenge. And people feel a lot of stress, hassle, learning how to use this new tool.

[00:09:03] And so the change management issue, the difficulty, is with specifically learning and laziness. That's what happens with traditional technology adoption. So when leaders are thinking about AI tools, they're thinking, well, we'll have people who are lazy, and we'll also have difficulties with learning these tools. And so that's what they focus on. And that's why AI pilots fail.

[00:09:29] Because leaders, IT leaders, CEOs, COOs, HR professionals, HR leaders, they focus on the wrong thing. They're focused on what they typically focus on when thinking about AI adoption. And that's not what you should focus on with AI tools. So here's why. AI tools are actually quite easy to learn. You don't use anything complex. You use natural language programming, meaning you talk to it. And you can ask it how to do things. And it will tell you how to do things.

[00:09:57] It's great at telling you how to do things and how to use it to adapt to your workflows. So it's actually surprisingly easy to learn. The change management difficulty, the resistance, so when I talk about my book, the subtitle is From Resistance to Results, it's not the technology that's a challenge. It's a psychology. So again, remember, it's not the technology. It's a psychology. For previous technologies, it was a technology. For this technology, it's a psychology.

[00:10:25] And there are three big blockers that I want to tell you about in particular that haven't really occurred with previous technology. So people are not threatened by their job. They don't feel a threat to their jobs by learning how to use a new CRM. But they feel that by learning how to use an AI tool and using it, they are training their own replacement because the AI tool is that good.

[00:10:51] And so they're worried about losing their jobs to the AI tool. Again, very different from previous technological developments. That's not something people were worried about when adopting previous technologies. So in chapter six of my book, The Psychology of AI Adoption from Resistance to Results, I talk about this psychographic profile. So there are a number of psychographic profiles relating to AI adoption. One of the biggest, probably the biggest that I've seen,

[00:11:19] is the AI alarmists. These are people who don't want to use AI because they're worried about their jobs. They might not come out and say this. We can talk about how to identify the AI alarmists. But that is the reason, the fundamental underlying reason, for their reluctance to use AI tools. And this, by the way, is based on over 100 consulting projects where I did hundreds of focus groups with many hundreds of participants

[00:11:47] and hundreds of surveys with many thousands, over 10,000 survey responses. That's where the peer-reviewed research for the book comes from. And so that's where I see this AI alarmist being a large group. So that's one, the fear and anxiety. Two is people who feel an identity threat. That's the big emotion for them. They're less concerned for their jobs. They might have some concern. But that's secondary. Their primary concern is identity threat.

[00:12:16] People don't feel a threat to their identity by having to look up information about a client in a new CRM versus their old Excel spreadsheet. But they feel threatened by a tool that can write a great email to this client, that can have marketing copy that moves this client to click, that can create images that impress this client, that can create great PowerPoints for this client. And so that's what they feel threatened by. They feel threatened to their professional identity.

[00:12:45] And their professional identity is very important to them as a professional and more broadly as a source of meaning in their lives more broadly. Our work to us, to many people, is an important critical source of meaning. So for people for whom work is a big source of meaning and who feel threatened to their professional identity, that's the pragmatic resistor psychographic profile. Again, same chapter six in my book. The last one.

[00:13:14] So these are the two big resisting groups who don't want to adopt AI on there individually. The last group is the reluctant adopters. Now, this psychographic profile that I want to share with you about, they adopt AI individually, but they don't talk about it to colleagues or their supervisors. You might have heard the term shadow AI usage. These are people who feel shame, social stigma around AI usage.

[00:13:44] They don't want to talk about how they're using AI to draft their emails, to create their PowerPoints, to write articles, create images, to create financial reports, to financial analysis, and so on. So many things that AI can do. It's an incredibly flexible tool. It's a great technology. The technology is great. But the social stigma is high. I'll tell you about an example. So a few days ago, I did a presentation for a Vistage group, which is a group of peer executives.

[00:14:13] There was something like eight executives in the group, mid-sized businesses. And when I started talking about shame, one of them told me that he just had an experience in this company like this. So he's a CEO. And the VP of sales did a big presentation that the VP of sales does every year on the best client that they got that year, the largest client. So it describes all the context, how they got them, all the stuff and the implications for the company

[00:14:43] and so on. Now, after the presentation, which went swimmingly, it went great, the VP of sales, and this is a presentation to the other executives, the VP of sales came to the CEO privately and said, hey, I want to tell you something. Previous years, this took me about three days to do, as you know. In this year, I delegated it to my executive assistant and he did it in one and a half hours using an AI tool, using Cloud. Now,

[00:15:13] I don't want you at this stage to tell anyone because other executives will think I'm cheating because I used Cloud to develop this presentation, which they thought went great. And so that's an example that's at the very top level of the leadership where shame and social stigma pervade AI usage. And of course, it percolates throughout the organization. And so people don't talk about how they're using AI.

[00:15:42] They might be using AI extensively, but then their additional time that they have freed up, they're using it to, you know, play around on Facebook and go to Amazon and buy things rather than putting it into productive activities for the company. And so as a result, the company doesn't see this productivity. So all three of these groups, the AI alarmists who feel anxiety and fear, the pragmatic resistors, the people who feel identity threat,

[00:16:12] and the reluctant adopters. These are the people who feel shame and social stigma. That's where it comes back to the 95% of AI pilots that don't scale, that fail to scale, that they don't show return on investment for the scale when people try to scale them up. Wow. Well, those are significant. So you're saying across those three cohorts, those are the three primary constituencies that fed into

[00:16:42] the 95% failure rate, right? That's right, Bob. So when you think about, I guess there's a lot of conversation over the last couple weeks around AI slop and sort of this backlash around, you know, AI's writing and now, I guess, detectors, I suppose, are getting better in theory about detecting it, but also we humans claim to be able

[00:17:11] to spot it more easily and more readily. So do you think for that third group, that shadow AI use, it may come down to, you know, specific use cases, but in your examples, it seems like we've got to unpack that group a little bit further to say some things people might be using their own AI for, could be like legitimate, you know, work-related things that they're comfortable

[00:17:41] offloading and otherwise it might be things that, where the expectation is that a human is putting, you know, words to paper as it were and taking on that particular task, right? So people may, I don't know, I think there's a lot of nuance to that third group, right? Because some people may think that there's, maybe some of that work is core to their job

[00:18:11] and it ties maybe to the second group, but some of it may just be, well, I'm not great at, I'm not great at PowerPoint, so why wouldn't, why wouldn't I, you know, augment my own, you know, sort of repertoire of skills and let AI, you know, help me with that. And I suppose there's an element of, you know, does any of that trigger newfound, you know, transparency or not? Do they stay in the shadows? Sure, so let's talk a little bit

[00:18:41] about the concept of AI work-slop. This is poor quality work that has been produced by AI and there was recent research published in Harvard Business Review conducted by the company BetterUp showing that something like an average of two hours a month is wasted by colleagues correcting poor AI output that has been sent to them by their own colleagues. So this is only internal. We don't know how much time has been

[00:19:10] and money has been lost externally by these people sending poor output to clients or something like that. And also, we know internally $186 is lost with such output per month. So again, two hours per employee per month and $186 per employee per month. And so these are problematic numbers, obviously. And so what's happening here? We have a combination of things. There are two things that I want to talk about. One is

[00:19:40] malicious compliance. Malicious compliance. And that happens with groups one and two. And I talk about this in chapters 7, 8, and 9 of the book. So what's happening with these groups. And I also talk in those chapters how to address each of those groups. But let's talk about the problems with the work slot. That's what you asked about. So with the AI alarmists and the pragmatic resistors, if they're forced to use AI, if you tell them you must use AI tools

[00:20:09] to do this, this is the new company policy, which definitely many companies do, what tends to happen is that they use the AI maliciously, meaning they just press the AI button and they get a crappy output, a sloppy output, and then they send it onward to their colleagues, to their clients, vendors, whoever. And then when their colleagues and their supervisor especially tells them that like, hey, this is bad quality output, why did you do this? They're like, well,

[00:20:40] you told me to use AI, I used AI, you know, that's it. That's what happened. And so that's malicious compliance where they're trying to specifically use AI badly in order to undermine AI adoption. They're like, well, I'm not going to be using AI if you don't like it. So that's malicious compliance. So that's one dynamic of what's going on. Then a second dynamic is people who use AI who aren't trained to use it. I've lost count

[00:21:09] of the number of companies that just take AI, they get, let's say, cloud subscription, which is the most popular AI tool for companies now. And cloud is great if you use it well. But if you don't know how to use it well, it might not be so great. So the company leader just is excited about using it and she knows how to use it well because she figures out how to use it and she's like, well, everyone needs this. And she just rolls it out to everyone in the company.

[00:21:39] She's like, hey, here, go ahead, use cloud, experiment, do things, make sure to use cloud, use it well, and so on. And then people start using it, especially junior people start using it because they're excited about the technology and they don't know how to evaluate the right standard because what happens, especially with junior people, is that they, as they're going through the process of creating an output, a deliverable, whether that's

[00:22:09] a marketing email, whether that's a financial analysis report, whether that's an article that they're writing or an image or marketing copy or whatever it might be, PowerPoint. They have a, they work out, they figure out, okay, I'm not sure about like what's going on here as I'm doing this. Let me go and ask Bob in who is the, you know, senior guy or Mary, you know, senior woman

[00:22:37] and try to figure it out. They get mentoring. They work out various elements of it during the process of creation. And so they don't know, they haven't been trained on using the AI tool effectively, which involves not, which involves creating the initial draft of whatever you're doing using the AI tool and then evaluating it against whatever the best standards are of the company and your own knowledge. So people who are experienced like you and I, Bob,

[00:23:07] we know what the output needs to be. And so we can know and evaluate the output pretty quickly and we can work with the AI tool even if we're not experts to get to that, change its initial draft to get that output. But junior people who don't know what the output is and who haven't been trained on how to work with an AI tool, they don't have that benefit. And so they produce sloppy work that's sent downward. And then I was,

[00:23:36] I remember having another conversation with that executive message group where somewhere another CEO told me that like, okay, I've lost count of the number of times where I had a junior person do a presentation and the presentation went great. And then I asked them for the reasoning behind certain recommendations in the PowerPoint they presented or the interpretation of it and they had no idea. And it's because

[00:24:05] the AI tool came up with something that looked good but the junior person didn't create it and they didn't evaluate it, they didn't really understand it. And so it wasn't, the output itself was not sloppy but the thinking of the junior person was sloppy. And so the combination itself was sloppy. You might write a great sales email but when you go into the actual meeting you're not able to handle it well.

[00:24:35] So that's part of the sloppiness that's going on. So you have those two elements happening. And so it's less to do with the reluctant, it's less to do with the reluctant adopters than those two elements. The malicious usage by the AI alarmists and the pragmatic resistors and junior people not being trained on it and not having the skills that more senior people have to do that. There's a reason for why junior people

[00:25:04] are being hired to lower rate in AI exposed industries. And so we can talk about impact on the job market it's definitely there. And so that's definitely happening. Yeah, I mean I think some of that tracks just in terms of you know a lot of times when I'm using AI it's sort of like I'm talking to an intern and you know less tenured employees well there's a reason we don't give them their own interns

[00:25:33] I guess. Because they don't they don't have the experience and the judgment and the you know tacit knowledge of how these processes work how the organization gets things done how knowledge is shared how conflicts get resolved you know those kinds of things and so they lack a lot of the context and they they'll miss some of the nuance that's necessary especially for you know maybe client communications or anything

[00:26:03] that's going to go up on a you know on a webinar or presentation for exactly the reasons that you pointed out but that they just won't even be able to it might sound good it might sound intelligent it might sound accurate but the second somebody asks a question to unpack it a little bit you know they're you know deer in headlights they can't explain explain you know what they've put forward and so so some of

[00:26:33] it is like I guess it ties back to your point before like yeah you told me to use AI and it sounded great it sounded better than what I could have put together so where did I go wrong but I also wonder if this ties into you know the gap that you've seen this sort of I think you may have referred to it as like a sort of split reality in terms of who has access and who is gaining the skills to effectively you know augment themselves with

[00:27:03] AI between you know sort of leadership who have that judgment experience etc and those who are just sort of you know hacking away at it and getting you know mediocre results which puts them right in that squarely back into the you know that that mid pile of sort of average output yeah and that's the problem that executives have motivation to use AI effectively

[00:27:32] because they are not worried about losing their jobs and they're not worried about identity threat they don't have enough time in the day to worry about threats to their identity and they are the ones who are making the decisions so they're not going to be finding themselves in either of the shadowy way but they're not using their freed up time

[00:28:02] for going on Facebook they're using their freed up time to do other priority tasks and so they are overloaded with tasks and so the executives don't have that problem of split reality and they don't understand where their employees are respondents are more excited and anxious and the rest

[00:28:32] are equally excited and anxious so you can imagine where the executives find themselves and where much of their employees find themselves so you do definitely have some people who are rank and file who are excited about AI because they're technology enthusiasts and they're excited and they want to make their work more efficient I book about the psychographic profiles you definitely do have some

[00:29:02] staff who do use AI and who are excited about it but they're a small contingent of the large majority of companies they're going to be a small contingent the AI the executives are going to be there whereas the employees generally don't have good incentives to use it and they don't have good motivation to use it when they're not approached how to turn them into the right direction

[00:29:32] that can be done but right now executives generally aren't doing it because they're approaching AI adoption from the old school technology adoption mentality yeah I was that's you're you're teeing up exactly where I was going to go because I feel like I know some people would say just focus on the sort of optimists focus on build up that groundswell and

[00:30:02] you'll get more and more people to convert but it is concerning when people are almost adversarial to potentially their own growth as well as the team and organizational success that they put forward so how do you establish some of the psychological safety and some of the them you're not going to convert

[00:30:31] everybody there will always be people that think you know what I'm teaching this new AI solution just far too much about what I know and the way things work around here that to your point I really am it really does seem like training my digital backfield and you can't there's nothing you can say that convince me otherwise right there's always going to be those people but I also feel like you can take

[00:31:01] advantage of some of the tools and hopefully training proper training that the company is offering you take advantage of that because it's going to give you more flexibility and more in-demand skills that if they do ultimately have to part ways with you for reasons beyond your control at least you're making a pair there absolutely so let's talk about each of those groups and

[00:31:31] how to reach each of those groups and again that's chapter 789 in my book so let's talk about the AI alarmists there are two things that need to be done to address the AI alarmists and again as you rightly point out Bob you won't be able to reach all of them and it's just a reality but you can reach many of them one thing you can do is make a commitment that if the company sees productivity gains from AI efficiency and effectiveness we can talk about metrics and what

[00:32:01] efficiency and effectiveness means later if the company sees gains and productivity from AI the company will focus on growth not cost cutting not reducing jobs so it's a commitment that large majority of companies can make not all companies but large majority of companies can make it I encourage companies I work with on AI adoption training and consulting to make that commitment makes it much easier to reach the AI alarmists and to the lesser extent other

[00:32:31] groups as well who other psychographic

[00:33:03] they need to their profit by 9.5 faster so they're basically increasing their profit faster than increasing their head count and so you can see where the efficiency is going they are increasing their profit by nine percent they're increasing their head count by 6.5 and so they're using less people than

[00:33:33] and they're taking market share away from companies that are not using AI effectively. And so the people who are losing their jobs are losing their jobs in the companies that aren't using AI effectively, because those companies are losing market share. So the companies that are gaining market share, that are using AI effectively, are growing their profits, and they're growing their headcount, but the headcount is growing slower than the profits. So that's one. So you have evidence behind that. So that's one thing you can do. And two,

[00:34:01] what you can do, and every company can make this commitment, is say that, hey, to the extent that you use AI effectively and you learned how to use AI effectively, your job here will be secure. And that's an accurate commitment, pretty much an accurate commitment, because who would you rather let go? Would you let go of somebody who is much more productive because they're using AI, or somebody who is not using AI and therefore less productive? Of course, you want to keep the person who's

[00:34:31] using AI. And as you pointed out, Bob, that person will also just, if the company, for some reason, outside of their employees' control, if the company fails, whatever, then that person will be much more safe and secure because they have AI skills. But focusing on internally what the company can control, yeah, you're not going to fire people who are using AI effectively. You're going to let go of people

[00:34:58] who aren't using AI effectively, of course, and who are less productive. And so that is basically flipping the script for the AI alarmists. Currently, they're thinking that, well, I'll be replaced when I, because I use AI and I teach it how to do all these things. What they need to be thinking and realizing and what the leadership, this is about leadership communication and company policy. And so this is something that's not that hard to do. It's about company call policy and leadership

[00:35:27] communication. You can make the commitment and get them to realize that they will lose their jobs if they don't learn how to use AI effectively. And they will lose their career because wherever they go in the future, they'll need to face AI usage and they'll need to be effective at it. And the people who definitely have a lot of research showing that people who are, have AI skills are much more in than people who don't. So that's the AI alarmists. So that's the first group. That's what you need

[00:35:56] to do with them. Now let's talk about group number two, the people who are the pragmatic resistors, who feel that identity threat. And so here we get into typical training on AI tools. The typical training on AI tools teaches people how to do the things that are central to their job, whether it's financial analysis, whether it's how to write good, great sales outreach emails, whether it's writing

[00:36:24] various reports, whatever it might be, doing tax preparation, whatever it might be. That's what the typical AI training involves. And you know what? That's very triggering to the people who feel an identity threat to those things because they feel proud of being able to do those things. So instead, when I do a training for a company, for their employees, for the rank and file employees, what I do is I do a survey on, hey, what are the things that you most hate doing in your job?

[00:36:53] What are the most annoying things? Where do you have to look up data across a variety of sources to create the packet that you need to then do the tax preparation or the financial analysis or writing the article? Nobody likes to do that. What are the things that you would most like to get off your plate? And that's what the training needs to focus on. You need to prepare custom prompts and agents that are going to be specifically focused around the things that people hate doing, they don't want to

[00:37:23] do. That's what you'll certainly not go as fast immediately as if you automate the tax preparation itself or whatever the financial analysis report itself, whatever is central to the role. And AI can certainly do that. But you go slow to go fast, you will actually get tax preparers and accountants

[00:37:47] and salespeople and HR people, all sorts of operations to actually adopt the AI tools because they want to. They want to get these things that they hate off their plate. They don't want to be looking up in various sources to prepare the packet for the things that they love to do. And so that's what you need to focus on. You need to focus your preparation and the tools that you provide to them and the training on the things that they hate. So that's an AI training that actually

[00:38:16] works. And that's an AI training that people will actually use afterward. So that's the second group. Let's talk about the third group, the people who are the reluctant adopters who are ashamed of it. Of course, the first thing to do to address that group is that leaders need to model AI usage. They need to model it. So that's the first group. So the leader, the VP of sales, I told the CEO that

[00:38:42] what he needs to do is go back to the company and tell the VP of sales to send an email to all the executives and tell them about how they use, how he used AI. His gave his time to his executive assistant, it wasn't even he to create this presentation and talk about it to his team and get this message spread widely. And the leaders need to model AI usage. They need to talk about it. They need to talk about

[00:39:08] how they're doing performance evaluations. Well, that combines something that leaders hate with an important activity that their subordinates care a lot about. So performance evaluation is pretty much on the lowest level of concern and care and passion for any leader. They really dislike this task. They know it's important, but they dislike it. And so if you have an AI tool, instead look over

[00:39:34] the past year of deliverables by your employees and create a report for the various deliverables, whether it's by email, whether it's from Trello, whether it's in SharePoint, whatever it might be, get those deliverables and evaluate them accurately, and then look at, combine them into a project. Then you're going to be giving them a much fairer and accurate performance evaluation,

[00:40:00] and you're going to spend much less time than the typical performance evaluation, which honestly only looks at the last month of performance, because that's what the leader remembers. And that's an example. That you can talk to your subordinates about. That, hey, I've done this performance evaluation using this tool. It's great. And so you need to model it. And so that's one. Second, training. We talked about what a good training looks like. So when you give them the training that's a good training,

[00:40:29] and they actually start using it for those things that they hate doing, well, the social stigma starts to disappear, because everyone agrees that, yes, this is a great usage for AI. Let's use AI to do these things that we hate. And so now you can talk about how you're using AI to do the things that you hate. And over time, of course, you will use it to do other things because it's so good, but now you already addressed some of the social stigma, and it expands. Now, in order to address

[00:40:57] it further, what you need to do after a good training is to set up an asynchronous channel where you communicate about new uses of AI, but the leader praises whoever communicates about these new uses of AI. So praise is very important. So whether it's going to be a Teams channel or a Slack channel or a Trello card or a Summit card, that's where you need to have the leader positively reinforce whatever people share

[00:41:25] about their AI usage and ideally provide them with rewards. So maybe a small bonus of some sort when they share a new prompt that others can use to make their work more efficient, something like that. Now, ideally, you'd also set up a synchronous way of talking about it. So for example, a weekly meeting sharing about your AI usage. I understand that not all companies can do this, especially ones

[00:41:51] that are fully remote. And so an asynchronous venue is one that everyone can do. And so I recommend at least having an asynchronous venue with reward and praise, and ideally, you can have a synchronous one. So those are the three things that you can do for that third group, the reluctant adopters who feel shame and stigma. And so again, that's chapter seven, eight, and nine in my book. And that's the way

[00:42:17] that you overcome each of these things. So again, a combination of effective leadership communication and policy is going to be for the AI alarmists. Then training on the things that people hate and adopting those, that's going to be for the identity threat. And then for the third one, the reluctant adopters for the social stigma and shame, that's going to be about leadership modeling, training, and communication from a synchronous channel with reward and praise by the leader.

[00:42:46] So I like your thought process on each of those on this, on the, well, I'll start with the third one. So for the third one, what's interesting is I've talked to some folks around this sort of shadow AI use, and it was almost like they were trying to, like they had this efficiency boost and they didn't

[00:43:12] want to, it's like not disclosing that you're taking like performance enhancing drugs, right? It's like they didn't want to let anyone know what their secret sauce was, or that they were using something, you know, something that wasn't sanctioned by the organization, but, but they felt it was providing better results. It was a, it was a either a faster model or it was just a, you know, a better

[00:43:40] model, or it was some tool that maybe it was just something that was still a Skunk Works, you know, project in, in beta and, and they had access to it somehow and they were trying to get ahead. So it sort of reminded me of, you know, before social media, people used to not share their expertise publicly, right? Because, well, that I won't be the go-to, you know, expert when everyone

[00:44:09] knows, you know, ties directly to that identity piece that you're talking about in the second group, like, well, I'm, I'm known as the expert on this, this topic. So if I share that, if I share all that, then I'm not the expert anymore. And then I'm not as in demand. And then, you know, they won't, if I threaten to leave, they won't make me a counter offer or, you know, convince me to say or whatever. So, so there's, there's definitely a lot of psychology that goes into, you know, why someone

[00:44:39] does what they do and, and selectively shares what they're doing. I also think in that group that, well, maybe this was in the, in the third group, but I was thinking about how the organization, you know, if you do share that, you know, what you're working on and where you've had success, you should also be sharing what didn't work, right? What do you need to watch out for, right? And so

[00:45:06] I think some of this is about not being selfish in a way, like, do you want the team to be successful? Because generally, you know, if you work as part of an organization, you're working as part of the team, maybe multiple teams, you're part of an immediate team, and then you'd have your department, your division, your line of business, you know, whatever. And so, so you have stakeholders and

[00:45:31] people that are relying on you to collaborate effectively all over the organization. And so, just feel like there's some level of sort of contributing to the betterment of, of the whole that needs to enter some people's psychology in a way. And so, I mean, you're, you know that neuroscience better than me for sure, but it just seems like some of this, and maybe it goes back to

[00:46:00] that communication, right? Like what's not just what's in it for you, but what's in it for you, the collective, you know, you. And, and I think that can bleed into, just to kind of go full circle, it kind of bleeds into how are we making better decisions, not necessarily as individuals, leadership or otherwise, but how do we, how do we get better at collective decision-making, collaborative decision-making? Because you're going to be in a, in a room with a bunch of

[00:46:28] decision-makers potentially that have to take all this, you know, evidence and all these perspectives and, and all that cognitive diversity and come up and make a better decision that affects a lot of people. I hear what you're saying. So I think here we're running into like what you're talking about, like that last one about like, okay, why are you not contributing to the team and so on? We're running into something that's again, going back to the split reality between the executives

[00:46:57] and the rank and file staff. So the executives feel a sense of ownership for the company. They feel a sense of ownership for the team, for team outcomes, for company outcomes. And if they've done a great job of building up a company culture, that percolates to all the employees. But let's be frank, it's hard

[00:47:22] to build a good company culture. And there are plenty of people who are much more interested in themselves than they are interested in the team. And so that's just the reality of the situation. And we need to face that. And so it's nice to hope for people to be communally oriented and oriented toward like the culture of the organization. But that's not something that should be expected. That's not something that

[00:47:51] we should expect to happen. So my book is really oriented toward people as they are, rather than toward people as we would like them to be. And realistically, people are much more, so everyone essentially will respond to individualized incentives. People who are communally oriented will also respond to individualized incentives, right? And people who are not communally oriented will also

[00:48:18] respond to individualized incentives. So that's why I prefer to encourage executives to base company policy and training and everything around AI adoption toward not assuming that people are individually oriented and care about the team. This policy, the stuff I'm talking about in the book, the psychology of AI adoption at work from resistance to results, is assuming the baseline of like ordinary workers,

[00:48:48] people who are nine to five, people who leave their jobs at home, people who leave their jobs at work, who go home, who are not like thinking so much about the team, who are thinking about what's in it for me. And that baseline assumption is going to be safe with anyone, whether they care about their jobs, like if they care about their jobs, you know, that's great. But if they care about their meaning, if they care about their teams, if they care about the company, that's great.

[00:49:17] And they can be motivated by that. But my approach is going to be toward going to people where, let's be honest, most are. And so most people, when you look at Gallup surveys, something like 63, 64 people are not, don't report being engaged in their work. And so let's target those people. Those are the people who we need to actually realistically orient toward.

[00:49:45] And so those are the people who I think we should focus on. And those are the people who's, for whom I'm providing a message that leaders can use to reach those people. I'm providing policies, I'm providing training, I'm providing guidance. And how do you reach those people? And those people are not going to be one simple psychographic profile. They're going to be the three distinct ones that I went through. So you need to reach all of them in a distinct way. But you can.

[00:50:13] And the crucial thing is leaders know that most people aren't engaged in their work. They know it. They know that most of their employees are treating it like a nine to five job. And they think that the change management difficulty they have to overcome is like previous technology. It's a difficult thing to learn. And they need time. And they don't want to go through the stress of learning it. And they're

[00:50:39] tied to their ways. But that's not what's a blocker. So that's the gap. That's the real gap that I've seen in all my consulting projects. So the leaders are not targeting the right change management problem. Do you think that on top of all the advice you've given here, and I know there's a lot more in the book. But there's part of it. I was thinking about your workshops, like experiential project-based

[00:51:07] workshops, right? And so it seems like that could be a very effective way to get people to realize just how powerful this can be for themselves. You're right. You got to deal with the hand that you're dealt, right? You can't just have these rose-colored glasses and think that, well, why can't

[00:51:30] everyone just hold hands and sing kumbaya? But it seems like sometimes you've got to, people don't know, they're just resistant because they feel like being resistant and they don't like change. But if you actually show them how to use it effectively and responsibly, I know you care very much about responsible use. And I think we both are proponents of being responsible by design, which means

[00:51:56] showing people how to do it effectively and responsibly, ethically, et cetera, can really yield all kinds of do-it-yourself folks that you didn't expect. So I don't know, it seems like that would be one way to sort of get these people to just do a 180 and really start to say, hey, you know what? This isn't so bad. Yeah, that's exactly right, Bob. So the crucial approach, as you identified,

[00:52:22] is to get people to start using the tools when they're trained on them and going back to what you want to train them on is the tasks that they dislike. And so with AI tools, giving them access to a broad platform like Claude, Anthropic or ChatGPT, Copilot, Gemini, and then in the workshops that I do, so again, to the survey, then I give them some examples of prompts and agents that they can create

[00:52:50] that are going to be focused on the kind of things that they identified in the surveys are the things that they dislike and they want to automate, they want to get off their plate. Then the first part of a workshop is going to be showing them how to use AI effectively with the examples of prompts and agents. The second part of the workshop is going to be them actually

[00:53:14] building these tools, and that's fundamental. They need to spend time building the tools. So I give them some skeleton tools, and then they spend time customizing it to their needs. Whether it's going to be, some people hate drafting emails, great, let's have them draft emails. Some people, they hate spending their time pulling information together from various materials to prepare the initial tax report they

[00:53:40] need to then work on to actually prepare the taxes for their clients. Or something that I recall from a workshop just that was about seven days ago for an insurance company that was going from Microsoft Microsoft of, I think, yeah, Microsoft GP to AP Flexi accounting. So they're doing major accounting transformation. So what they need to do is for each year, they need to export all the invoices in

[00:54:09] one spreadsheet format and then put it by hand into another spreadsheet format to import into the new accounting software. And they, in five days, it took them five days to do a previous year. Now during the workshop in 30 minutes, they were able to do another year for something that took them five days to do last year. That's not something they enjoy doing. They enjoy doing the financial analysis and the

[00:54:34] reporting, but nobody enjoys taking information by hand from one accounting system and a spreadsheet format and putting it into another accounting system, spreadsheet format. So those are the kinds of things that people need to learn and adapt and build themselves and use it for the activities that they hate. And so that's the DIY approach, where people learn how to use AI to improve their own workflows.

[00:54:59] You don't need to buy like complex AI products. The large majority of AI products can be built by people. They just start with the tasks and the tools that they dislike, but there's no reason why claims letter agent, why you need to buy an AI tool for an insurance company to do claims letters, which insurance companies are buying all the time. No, you can build your own claims letters agent and have your

[00:55:25] people build this tool that then they can use extensively and thoroughly throughout the company. And so these are the kinds of things that can be built effectively and successfully DIY by your employees when you have the right basis. And when you flip the script for them and you address those three groups effectively. Yeah, no, I think that's awesome. The, the, you hit on a couple of the points that we didn't,

[00:55:54] I think I forgot to ask you before, but it was really, you're really emphasizing, first of all, let's start with business challenges. Let's not start with technology. What, what, where's the friction? Where are the things that humans should not do? So don't say, what can AI do? What can it, what, what should stay on a human beings plate? And then what can we offer to AI that we don't like to do,

[00:56:19] or it's, you know, cost-effective or it's dangerous or it's repetitive or whatever it is, right? And you're also in the process of doing that. You're actually giving these employees, this workforce, you know, their sort of agency back to actually now build something the way they would love that their other software, their legacy software worked, right? And now they have more autonomy and more control

[00:56:49] and they can understand how it's all built and things like that. So there's just so many positives if you get people in the right sort of frame of mind. Yeah. Absolutely, bro. Absolutely. Love it. So Dr. Gleb, this has been a fascinating conversation. I want to be respectful of your time. Definitely going to include your link to you, your site and your new book in the show notes. Congrats again on the book. Any parting thoughts as we wrap up?

[00:57:19] I just want to mention to people that if they want to get a free sample of the book, introduction and chapter seven, they can go to my website, disasteravoidanceexperts.com forward slash AI book. Again, free sample of the book, disasteravoidanceexperts.com forward slash AI book. If they already bought the book, whether they pre-ordered it or they bought it in a physical bookstore or they bought it in Amazon or Barnes & Noble online or in person,

[00:57:46] just take the receipt number and go to that same website, disasteravoidanceexperts.com forward slash AI book for a free assessment on how to adopt AI effectively in your company and a manual on the seven critical mistakes that leaders make in AI adoption. Again, that's going to be disasteravoidanceexperts.com forward slash AI book. Dr. Gleb, thank you again so much for spending some quality time with me. It's been really,

[00:58:11] really intriguing discussion. Thank you for all you do to help organizations adopt AI responsibly in a human-centric way. I really appreciate it. Thank you for a great conversation. It was a pleasure to be on. Dr. Gleb. All right. My pleasure. Thank you everyone for listening. We will see you next time.