Using AI To Build Leadership Skills with Gary Lloyd

Gary Lloyd is a change leadership practitioner, author of Gardeners Not Mechanics: How to Cultivate Change at Work, and creator of Leadership Skills Lab and Leadership Skill Builder. He has spent much of his career leading complex change and developing leaders, including more than a decade as a mentor and former co-chair of Warwick Business School’s mentoring programme. His current work explores how AI can support safe practice, reflection, and genuinely personalised leadership development.

A Few Quotes From This Episode

• “People are more open to feedback from the AI on the conversation they've had...it feels less threatening, it feels less like a human being."
• "You can read about it till the cows come home...if people want to know what the possibilities are for them reinventing the way that they deliver leadership training, then go and get your hands dirty a bit. It's unbelievably easy."

Resources

• leadershipskillbuilder.ai

About The International Leadership Association (ILA)

• The ILA was created in 1999 to bring together professionals interested in studying, practicing, and teaching leadership. Attend The Global Conference in Toronto, October 28-31.

About  Scott J. Allen

• Website
• Weekly Newsletter: Practical Wisdom for Leaders

My Approach to Hosting

• The views of my guests do not constitute "truth." Nor do they reflect my personal views in some instances. However, they are views to consider, and I hope they help you clarify your perspective. Nothing can replace your reflection, research, and exploration of the topic.

♻️ Please share with others and follow/subscribe to the podcast!
⭐️ Please leave a review on Apple, Spotify, or your platform of choice.
➡️ Follow me on LinkedIn for more on leadership, communication, and tech.
📜 Subscribe to my weekly newsletter featuring four hand-picked articles.
🌎 You can learn more about my work on my Website.

Transcript

Copy Transcript Copied

Welcome To Practical Wisdom

Scott Allen 0:00

Thanks for checking in from wherever you are in the world. Welcome to Practical Wisdom for Leaders, the official podcast of the International Leadership Association. I'm your host, Skim J. And each week I invite executives and scholars to explore the real work of leadership. This podcast is part of a global community committed to helping leaders run. If you find these conversations valuable, connect with me on LinkedIn, subscribe to the podcast, and share this episode with others. To learn more about me, visit www.scotjmlin.net. And now, here's today's episode. Okay, everybody, welcome to the podcast. Thanks so much for checking in wherever you are in the world. Of course, practical wisdom for leaders, the official podcast of the International Leadership Association. Today I have a returning guest, and I'm super excited about this conversation because I always love checking in with Gary Lloyd. And Gary and I have done a couple podcast episodes. This might be number four, Gary. And yeah, something I'm trying to catch up in terms of some reams. Yeah. You know, he had a when he approached me, he was like, I've got about 12 episodes. And so yeah, we kind of moved

AI Slop And Tool Misuse

Scott Allen 1:15

through those, but that was a lot of fun. We got a little time on that one. We got a little work to do, but there's plenty to talk about. And, you know, of course, you have been on this kind of experimentation, explore exploration um initiative in the last few years of really, you know, again, building and experimenting and exploring with artificial intelligence. And then how does that intersect with leadership education? And so I think I'm just really excited to kind of hear what you're seeing. What are two or three observations you have? You know, this topic has been on my mind. I more and more, Gary, I'm kind of bumping into, and you know, I've been exploring LinkedIn a little bit in the last year more intensely. You've got these folks who's their their expertise is is Claude. Uh I keep bumping into managers that are basically using Claude to manage or trying to use Claude as a default to manage, or consultancies that are overusing Claude to try and and nothing replaces expertise. Like it's a tool, but it's not your expertise. So I think I I don't know. I mean, I'm I'm seeing some abuses of some of these tools out there. Maybe we'll get there, maybe we won't. But I'm just excited to hear about what you're seeing as you're building. So uh two or three things. What's on your radar right now?

Gary Lloyd 2:35

Well, well, a quick as quick aside, I don't know if it's true, but something dropped into my inbox just now that um said that LinkedIn are introducing an this is AI slop button that you can you can you can vote on. I like that. I like that. That's good. So so everybody needs to be cautious. Um how how good it is or not, I don't know. We'll get into the uh into the work thing. I uh uh fortunately you warned me of this this question, Scott, because I'd love to be really ultra fluent and just speak off the top of my head. But but actually I've had time to think about this. Yeah. And um well, I kind of had 12 hours to think about it. Um the so my three observations, which I've written down so I don't uh don't forget them, um, is um three observations of this. Uh one of the strongest thing that comes out from all of these podcasts is the knowing doing gap. Yes. I mean that just kind it just comes up time and time again. I won't pretend I don't shout uh uh uh at the iPod uh at the uh the recording when some of these come up because I just keep shouting AI, AI, AI. So the the main thing I've learned since we last spoke about AI, the technology has moved on quite quite a lot. And you know I I created something about three months ago, which we we can get around to talking to if it's of interest to people, um, which is that it's now not difficult to create learning experiences in which practice supported by a tutor or not and scaffolding uh are a real possibility now. So, in terms of knowing and doing, um, it's really not difficult to put together learning experiences where people practice doing in a safe environment. So that's that's observation uh number one. Okay, um observation number two is kind of related, um, which is it's it's it the technology of talking to a computer to get it to create an app for you

Practice To Bridge Knowing Doing

Speaker 1 4:55

is now easier than ever. I shouldn't shouldn't say easier than ever, it's really easy. Um I I I created something for Doris. Doris Doris uh is an artist and she wanted to organize her work and stuff like that. And she'd she'd experimented with every single app that was out there, but it wouldn't do exactly what she wanted. Because we all know, any of us that have been around technology at all, know that if you're gonna buy buy some software, then what you have to do is to bend your processes to fit the software to get the best out of it. Otherwise, you'll spend a huge amount of my money on customization. Yeah, that is no longer the case. So Doris in Frustration came to me with a promise I'd made to her long before, and she said, Could you build me an app? Two hours later, there's her app. Now it's not that we didn't iterate around it after that, but she now has something that works in exactly the way that she wants it to. Well, and this is this is true for leadership educators as well. There's there's no reason why leadership educators, there's no reason to do a double negative now. Leadership educators now have access to the technology. And believe me, it's easy. When I say it's easy, it's really easy. You know, Scott, I'm not a technologist. Uh, my background is uh is changed leadership, and I kind of stumbled uh in into this territory just uh experimenting through uh curiosity. Absolutely bespoke, this is exactly what I need. You can get by talking to the computer now. So that's observation number two. Learn creating learning experiences that close the gap between knowing and doing by putting doing in front of the people. And the third one is really something's been bugging me. You touched on it in the intro, something's been bugging me for I don't know, three to six months uh now. Um, and I've only it's taken me this time really to wrap my mind around what's the lesson, and that's the rise of agents. Now, if if we go to the social media channels, then agents are everywhere. Yes. Um and they're great, but they introduce a lot of risks because in all the conversations people are having about agents, they're not talking about what we used to call hallucinations, or we still call hallucinations, but but are better described as mistakes, that they make mistakes. Yeah, and if an agent is a number of does a number of steps, then each step is an opportunity to compound uh a mistake. So in terms of the the observations uh around that is I've just started to see that the big AI labs at the corporate level, at the organizational level, are kind of realizing now that this is a big risk for organizations. So they're starting to build uh frameworks for risk management, essentially. So what I'm trying to drive towards here is saying that you know, from our point of view as leadership educators, if we're building tools for ourselves, we need to be aware of risk. And if we're delegating off to something to do work for us, we need to treat it like a human being. You know, we need to say what are the limits, what are the discretionary limits? What's what's the audit trail look like? How did you come to that conclusion? So we need to build these things in, and organizations are starting to realize that they uh they build uh to build these things in. And I think that it's kind of it's a challenge that the the people were educating. I say we're educating, you're educating the the the the listeners to this podcast leadership, doing development. I mean, you and I have talked before about uh you said like a uh you thought that a key leadership competency was um uh technology literacy or the kind of morphing into AI literacy. And I had always kind of gone, yeah, I'm not yeah, I'm not so sure about about that because it's very specific. But at this specific in time, time uh spacing time, I was talking earlier in the week to a mutual friend of ours, Gary Strike, uh, about the role the the role of uh CEOs. And um, you know, even people that are at CEO levels, top of their their list of worries worries, top of their list of concerns at the moment should be risk uh risk management and managing an uncertainty.

Build Bespoke Apps By Chatting

Speaker 1 9:40

That may that may change over the next 12 months or 24 months as the systems change. But you know, it was I think about three years ago, Sam Altman, who heads up um OpenAI, said in a year's time nobody's gonna be talking about uh hallucination. Uh 18 months' time, he said nobody's gonna be talking about hallucination. And he was right, but he was right for the wrong reasons because what has transpired is that it's a feature, not a bug. And so you just like uncertainty in organizational life. Um, you know, I wrote about this in in in in my book, which is you know, and we need to work with the knowledge that uncertainty exists and how we work in uncertainty and not trying to control it. It's it's it's it's a feature, and it's disappeared because the people I saw that Tata Consultants, sorry, going off a bit tangent here, uh Scott, which is uh Tart Tata Consultancy Services, which is a big IT provider out of uh India. Um they they're training like 8,000 people in to be AI engineers to help their clients. And I kind of go, well, what the hell are they what are you doing? What are 8,000 people doing out there? It's absolutely astonishing. So there's a massive push towards using AI, uh, and people are being measured on it, and it turns out to be very expensive. And what AI can deliver is extraordinary, absolutely extraordinary. But the risks need to be managed. Yeah, so those are my three things. I've got another way of answering. Can I have another way of answering it? Yeah, please, please.

Scott Allen 11:23

Yeah, let's let's let's you're gonna you're gonna add three more, aren't you? You're just gonna do six. I'm I've got four. Okay.

Gary Lloyd 11:29

I've got four. No, it's just a different one, it's just different, different way of spinning it. No, I love it. I love it. Let's hear it. Different way of spinning it. It's my other notes, which is what are the questions that that we should be asking? So the questions that for me are how for leaders, how is the context changing for leaders? What does that mean for the skills that they need to manage that context? How do we help them develop those skills? And can our AI help us to help them develop those skills? So when I think about leadership development, that's kind of how I think about it. And when we when we kind of get to the the last two of those uh four, I guess that's where the most of my time has gone where uh over the last probably three months or so, which is with starting off with some high prof hypotheses, um, and then trying to set out to to prove them. And those hypotheses were um, can you evaluate how good lead people's leadership skills are, their thinking skills around leadership, through a deep conversation with AI, uh, and evaluate that. And I think I think I've ticked the box on that. Wow. And uh can you then use that to create learning experiences where this is where I go to Theo Dawson here. I mean, Theo says uh a skill is something you can practice, a leadership skill is something you practice, and a micro skill is something you can practice in the moment. Well, certainly AIA allows you to practice in the in the moment. And the thing I have now is it does it does the uh it does the assessment and then it gives a a menu of micro skills that says, Oh, what do you want to practice for the micro skills? And then it practice then it presents you with some scaffolding. You then go into uh um a supported tutorial, and then it gives you feedback on it, and then you just keep cycling around. And every conversation uh learns from the previous conversation. That that that wasn't difficult to build, but it's an illustration of the sort of learning experiences the AI now allows us to create. So that was a that was a longish answer.

Scott Allen 13:54

Well, there's a lot in here, a lot of really good stuff in here. I mean, I think uh that and you've been writing about some of this on LinkedIn in recent times. You know, it it might be is is AI the tool that's appropriate for the situation that we're in? Is is it an appropriate intervention? And to your point, I think at least in that point, in that post, you wrote, hey, look, you know, everyone's adopting and spending millions and millions of dollars and and and they've got FOMO, but is there a real use case for how AI is gonna help our customers or help the client that we're engaging with? And and so so what is that use case that's that's truly the need, right? And there's a lot of people with

Agents Raise New Risk Questions

Scott Allen 14:40

uh things that it can do, but is that actually the tool that's appropriate for the situation? So I think there's that piece that I kind of have on mind. There's another piece, this this knowing doing gap. And I think that's it's such a fun conversation in leadership education because I I just the other day I was with a a group, and of course, they all know that they should communicate effectively with one another. They should be aware of their nonverbals, but it was amazing to watch as soon as the conversation got a little bit heated, their nonverbals said everything, and they weren't aware any longer of kind of how they were showing up in the conversation. And so, you know, that that skill building element, that actual, can you actually, you know, and you and I have talked about this way too many times. You know, you can put me in a course on soccer and talk about it, but that doesn't mean I'm gonna be able to play football and I'm gonna be skilled at the sport of. And and so that is a conversation that I'd kind of love to think about and how can AI bridge some of that gap? I guess I wonder where you see this going from just specifically in leadership education. What are the possibilities that you see right now? You know, you've built this tool, some folks are using it, they're practicing with it. Um what are you seeing as people are using the tool?

Gary Lloyd 16:13

I think people are surprised by how realistic it is to start with. Um a couple of companies have avatars that would look much like we do now and have conversations for that. I haven't seen the complexity of avatars adding much to the conversation. People often like to uh just have the conversation. Um I think people are surprised by the feedback that they get. Sometimes they're upset by the feedback they get, sometimes they're pleasantly um surprised. But I find that people this is a surprising finding, Scott, that um people are more open to feedback from the AI on the conversation they've had. Bear in mind the thing that I've built is um well, both the experiments that I built were voice-based.

Scott Allen 17:07

Yeah.

Gary Lloyd 17:08

So people were talking as we're we're talking now, although they could they could they could that they they can't see a face. Um uh and people very um the way the it turned out to be the it seemed to be the best way to do this was to have the practice conversation or the evaluation conversation as a voice conversation, it's real.

Scott Allen 17:29

Okay.

Gary Lloyd 17:30

People sometimes report they feel under pressure, even though the the AI will wait forever before it before you to to to resp to respond. They still feel the anxiety, like I did earlier on when I couldn't remember my third point. Um you go, oh crack it. Well, I look at my notes, what will I do here? Um, so um, but they like written feedback.

Scott Allen 17:56

Yeah.

Gary Lloyd 17:56

Oh interesting. They they they like the feedback as text rather than somebody talking to them for some reason. We it's just the way we're wired as human beings, it feels less threatening, it feels less like a human being, and it feels because it what it's doing is it's switching from being a human being, which is having and by the way, the assessment conversations, I personally find them very tough. You know, it assesses me against uh five um what do you call them? Not heuristics, uh five uh principles or or yeah, yeah. Um there's a uh elements or they all come back to me in dimensions in a moment. No, the scoring thing, what's the scoring thing? A rubric, a rubric, yeah. Rubric, rubric. Thank God. Um so it scores against the rubric, five rubrics, um, zero to five out out of ten. And uh I so I set that up to be it's got to be absolutely neutral, it have offers no help or support. It doesn't, it doesn't um it doesn't go, oh that's a good answer. Yes. And it goes, so so why do you think that I mean one of the things about that experiment, going back to three observations, Scott, is that the thing that one of the things I really wanted to test is that uh could the AI um do more than say, oh, that's the right answer? Could it go, okay, why did why do you think that? What are your uh what are your underlying assumptions here and start to dig in? What we're trying to do here is to understand how the person thinks and evaluate their thinking process, not evaluate their answers.

Scott Allen 19:30

Does that make sense? It does. It does. I mean, I know Jonathan is kind of exploring and experimenting with a similar approach, but in adult development, right? Can you can you kind of assess the mental complexity of someone's answers and how they're thinking about a given topic? I mean, it's it's so interesting what you both are exploring, just in a little bit of a different domain. But yeah, it makes perfect sense.

Gary Lloyd 19:57

So so that so that bit feels personal to people. The voice based bit feels personal, but then when they they'll get the feedback and they'll go, okay, um I'm not that keen on the feedback, but you know, I'll it's I'm just gonna pick off this micro skill and and it gives a bit of scaffolding for the micro skill. And then they have a conversation with the supportive tutor, and then they get feedback on that at the end. And the feedback they read is kind of neutral because they know it's produced by computer.

Scott Allen 20:28

Yeah.

Gary Lloyd 20:28

Yeah. You know, and and and the the the I know this is weird, but the more accepting of of feedback that's presented in that way. Strange, strange but true.

Scott Allen 20:39

Yeah, I

Assess Then Practice Micro Skills

Scott Allen 20:40

mean, it's it's super interesting. I I did an episode, you would have listened to it with David Day, and we were talking about your your AI manager and what AI management would look like. And I just saw that Facebook came out with this little button or Meta came out with this little button that basically, and I think probably your glasses can do it as well, but they just record everything. And I don't want Mark Zuckerberg recording everything at work, but it is an interesting kind of thought experiment where just in the flow of my day, I could at the end of the day get data on how optimistic was I? How how what did I say that could have been interpreted incorrectly? What did I say that maybe uh stalled conversation? There's this episode of Black Mirror. Yeah, I think it was called the Entire History of You. Yes, yes. And you know, you could replay and go back to the meeting and people's eye contact. I I'm not looking for that.

Gary Lloyd 21:41

So so you could you could you could you could replay a conversation you've had with Jessica and go, no, no, no, I was right. Yes, this is this this is what I said. That sounds really good. The AI said I'm right. It's I'm correct. Look, let me replay it for you and go.

Scott Allen 21:58

See, it's that It's that nonverbal right there, that's right here that set me up. No, I I mean it it obviously we can go to some dystopian places, but I think there is some real learning that can occur as well. And I think it's super interesting. You know, it it it does feel different when the AI says, look, kind of uh you spoke 70% of that meeting. A consideration is that you leave some space for other people to share their ideas. I mean, there's real good solid learning right there.

Gary Lloyd 22:30

Well, you experimented with something around this, didn't you?

Scott Allen 22:32

A little bit, a little bit, yeah. But I think it's it's but not necessarily in the context of leadership. That was more kind of just presentations and communication. But yes, I mean there's a there's a program called Udly, which people could look to. But so I think there's there's a lot of opportunity. But again, in your space, it's interesting that learning that they want the written feedback and that the feedback coming from the AI, it's almost depersonalized. And I almost it almost gives me agency to decide what I want to take in and what I don't want to take in. I don't know. It it's that would be so fascinating.

Gary Lloyd 23:09

I think it's not uh I think it's not um I'm trying to think of a better word than shame. Um, but it's it's that kind of meaning which is when when you get the written feedback, you don't you're not embarrassed. Embarrassed is uh is a better word. You don't feel embarrassed because you've been you've failed, yeah, because you've been an idiot. Because so much of our schooling and so much of our education systems right across the most of the world is about uh not failing. It's about I won't say winning, but it's about passing. We're always passing exams and stuff like that. And if you don't do so well, you know, you've got 50%, somebody will go, uh shame, but you know, everybody likes you or something like that. You know, we're it all through our schooling, it's so exam-based that we're used to saying, here's a bar, get over the bar. If you don't get over the bar, there's something wrong that you you need to do. And that feedback is not done in a coachy way, you know, unless unless you're Carol Dweck, they're not giving you that feedback in a coachy way. There's one other thing that you said, Scott, that uh I wanted to pick up on, if I may, um, which was that and it this is really I think this is really, really important for uh leadership educators, um, which is and it's true in corporate life or in organizational life, uh, but it's as true for leadership educators, which is this is uh you know, when we had our last conversation about AI probably at the International Leadership Association Summit thing, which was what 18 months ago, I think.

Scott Allen 24:49

Sure, sure.

Gary Lloyd 24:50

Um the the yeah, I was singing this song and I'm still singing this uh song. I've I've read a and this is not anecdotal. I read the research that comes out from people often like Wharton, MIT, um the the good universities, the business schools, uh people are still using AI primarily to augment or automate processes or augment or automate tasks within processes to help them produce I nearly said slop to help them produce output that somebody else has to consume, by the way. Yeah. This is for I'm gonna go quick aside here, which is that um I saw somebody the other day saying they'd created an agent and it it reads and prioritizes their their emails and then it it writes the responses and it they they trust it so much now um that it sends the emails off. Wow. Uh as well. Well, it's probably somebody at the other end who has a similar agent, and the two agents are sending emails emails together.

Scott Allen 26:04

Meanwhile, both humans have no clue what they ordered for lunch.

Gary Lloyd 26:08

Speaking of Wharton, here's here's something for your jargon dictionary. His hang spot on your jargon bingo card. Yeah. I've I've had to write it down so I don't forget it. So speaking of Wharton, two um two researchers at Wharton, they did some research, uh Stephen Shaw and Gideon Knave, and uh they came up with what they're seeing is what they call cognitive surrender.

Scott Allen 26:29

Yeah, yeah. I I think that's that that's what I was saying at the very, very beginning. I mean, at least at least it sounds like what I was talking about, where I'm just outsourcing all of my thinking. And again, I might not even have the requisite expertise to understand what follow-up question to ask, how to go a little bit deeper, or is this good output? And I think again, that that it's becoming blatantly clear when someone doesn't have that knowledge or that expertise as a consultant, as a as a leadership quote unquote expert, as a coach, whatever it is, um they're

Voice Realism And Written Feedback

Scott Allen 27:10

just surrendering any of their own thinking. And then I I I have in my sphere Oh, sorry, I'm gonna Scott, I'm gonna interrupt you if you don't mind.

Gary Lloyd 27:19

Yeah, the it uh one of the things that that research showed was the uh I think a lot of it was done amongst lawyers. Yeah. Um, it's not just they don't have the expertise, they don't have the time.

Scott Allen 27:31

So they're just giving it away to it, and then um oh well, that's interesting as well. So go ahead and define cognitive surrender.

Gary Lloyd 27:42

Oh, well, I I I'm glad you asked me that because the researchers at Wharton have already done that for me. And they they have they have four components to it. Uh automation bias. Okay. Now, what they mean by that is if it's been giving you the right answer for a while, or an answer that you like for a while, your trust level goes up. Okay, this is the email emailing each other thing. Cognitive offloading, which is not not doing the thinking first, just offloading. And you know, we know ask an LLM and it will just say generate pages of stuff, even when you ask it not to, and you go, Oh, okay, it's pr it's kind of all right, it's good, it's it's sufficient. Yes. Um, the other one is uh conf confirmation bias, which is that because AI uh large language models we're talking about here, really, in the main, because large language models um get to understand us these days, um, this is I suppose it's kind of uh tied to the sycophanti one, but even when you go deeper in the obvious sicker fancy, it's but it's likely to say that even if you've been talking to it a bit verbally for a while, bat batting backwards and forwards, in the end, in the end, it's gonna it's more likely to confirm what you think. Confirm the the way that you think about it is right.

Scott Allen 29:04

Great idea, Scott.

Gary Lloyd 29:06

Yeah, yeah. And it would but it's it's much more subtle than that because you know they're uh they're they're persuasive, they know how to talk to us, they've learned what our buttons are.

Scott Allen 29:17

Now and they want us to stay on. I mean, I think people are not understanding this, just like social media, it's in Claude Anthropics best interest or or open AI. They're trying to keep us on the platform, correct?

Gary Lloyd 29:35

So let's make yeah, so let's make let's make a a quick correction here. You and I are both enthusiasts for AI. Oh, yes. So so so people uh people need to take our comments about some of the reservations in the context that that's as we want people to use them responsibly, uh responsibly. I've got one.

Scott Allen 29:56

Yeah, we have one more, one more of those four.

Gary Lloyd 29:59

One quote, and then I'm gonna come to the main point I wanted to uh to make. Well, we still have the fourth, we still have the fourth uh element of the the fourth one was false consensus.

Scott Allen 30:08

Okay.

Gary Lloyd 30:09

Um the uh the it's kind of uh it's very much like confirmation bias, but um it might tell you that this is the agreed view of this okay when it's not, it's telling you that it's the agreed view of it because it knows from the history that that's what your view is. I'll give you shall shall I give you an example? Sure. Uh I'll give you give if we've got time. Um I in preparation for this podcast, uh, I was gonna say, oh, if you go and ask it about um change uh leadership, it's gonna come up with Cotter and his seven steps or whatever of change change management. So I asked it and it didn't come up with it. And so I said, Well, I'm a bit surprised that you've you've you've passed the test that you didn't you didn't come up with it because it's the most common thing that is that is out there, and uh it said, Yeah, but essentially it said, Yeah, it's because I've got to know you and I've read your book. Oh um, so I've I've the answer is uh in that way. So so one big point I wanted to come back to uh with leadership educators and the fact that the this technology, despite the reservations, if we know how to use it, is absolutely phenomenal. What what I see people doing still is augmenting and automating work they do rather than thinking about what's the need, and how can how can we use this incredible technology to fulfill that need in a different way? And if the lead is if the need is to help people be better at a particular leadership

Cognitive Surrender And Bias Loops

Gary Lloyd 31:55

skill, and I tend to think about this is a conversation I had with the other Gary. I I I tend to think about and this is this is a bias on my part, I tend to think in uh studentship as a um a parcel, a group of skills. That's how I view it. Skill skills skills that skills that can be uh practiced. So the question for educators, I think, leadership educators, are go back to what I said earlier on, which is what skills do the uh people going to need now? And how do we prepare them? How can AI help us? That's the that's the big question, and that's the thing that I that I've been working on. And I think actually the technology to do that now is very mature. And you know, the technology is still moving very quickly around agents and stuff like that. But the the the the stuff about using uh large language models to do role plays and structure. I mean, what what has happened now with the so-called vibe coding thing, which is a thing like I did for Doris, where I just created an app for her, you know, any leadership educator now could go off to a weapon of their choice. Mine has become uh codecs because I had some mishaps with uh the Google ones. Like I was I was in Tuscany and uh got an email saying um we you've you we've just suspended your product uh project because you've you've blown your budget twice. So I think that what I would say to leadership leadership educators is that there is now something, if you've got chat GPT, if you've if you're using uh Claude, you're probably more than familiar with uh ChatGPT. Um but if you go to uh Chat GPT or Claude work and just ask it using plain English to create a role play for you or to create an app that does role play uh for you on a skill that you want your students or the people that you're coaching and mentoring to practice, just go try to build something and find out how easy it is, because this kind of goes back to the very first conversation we had on this topic, uh Scott, which is there's you can read about it till the cows come home. You know, that this is like your analogy with the dentists and practicing it. If people want to know what the possibilities are for them reinventing the way that they deliver leadership training or leadership education to the people that they support, then go and get your hands dirty a bit. It's it's it's unbelievably easy.

Scott Allen 34:31

Yeah. And I think that's the you know, I oftentimes end with what's the practical wisdom in this conversation? I think uh, you know, unless you have something better than that, please say it. But I think that's at least one of the pieces of practical wisdom here is are you engaging with the the technology? Are you using it? Are you setting up assignments where your students are developing that role play, where they're learning and you're kind of learning through their learning? I think there's a lot of opportunity here, but for me, there's and and you said it beautifully, you know, we are both enthusiasts. I think you can consume the dystopian, you can consume the utopian. I tend to consume both. You know, we've both read the coming wave, and there's some very significant downsides of this technology. And humans, since the stick or since fire, have weaponized every tool we've ever invented.

Gary Lloyd 35:32

I have I have a I have a great quote for you uh from somebody called uh uh Nenand uh Tomashev. He's a leading researcher, uh um Google Deep Mind. Yeah um and uh he said as soon as you switch off, you're rolling the dice.

Scott Allen 35:50

Yeah, yeah.

Gary Lloyd 35:51

Talking about the use of AI. And he's a I mean he's somebody that's developing this this stuff.

Scott Allen 35:55

Well, I just I just watched with Stephen Bartlett. I'll I'll send you and I'll put it in the show notes for listeners. But Stephen Bartlett had the the former, he was a futurist at OpenAI, and he left two years ago because he didn't believe in what they were

Get Hands Dirty And Beta Test

Scott Allen 36:10

doing and and and you know, his predictions of where all of this goes. Again, they tend to be a little bit more on the dystopian side, but I think we need to have eyes wide open and and be intentional. And I think that's as a leadership educator and how we use this tool, but also just on the the the the very grand 50,000-foot view of this topic, it's uh there's some there's some definite shadows that we have to be aware of.

Gary Lloyd 36:37

The learning system I was talking about is called the leadership uh um skill uh uh the leadership skill builder. I'm getting mixed up with my other one, the leadership skill builder. And my offer is this, which is if there's somebody out there that's got a group of students at whatever level, undergraduates to to postgraduates, um, if they'd like to work with me on beta testing that, then uh I'm really happy to support them in that because we would both learn uh uh a lot from doing it. So um, yeah, if you've got some cohorts of students and you think it might be useful, and I think the benefit of people uh doing this would be that without having to do what I said, which is go off and get over the hurdle of playing with something like Codex, they'd see they'd see live, see live what the possibilities are. And it might be that what I've built is absolutely not for them, yeah. But they might go, oh, you know what? So just just ping me up and um, you know, I'm happy to come in, talk to the students. Um, all free of charge, it's just an experiment.

Scott Allen 37:50

Well, you are doing uh a great service for the community, sir, and I always love our conversations. I appreciate you. And uh, we will, of course, do it again because this is an ongoing conversation of Gary Lloyd's adventures. It's the series, we're in it for the long haul. Uh so thank you so much, Gary, for the time today. I really, really appreciate it. Thanks for the good work that you do, and you know what? Have a wonderful weekend, sir.

Gary Lloyd 38:17

A great pleasure, Scott. Thank you very much. Be welcome to the family.

Scott Allen 38:20

Okay, okay.

Gary Lloyd 38:22

Okay.

Scott Allen 38:23

Thank you so much for listening to Practical

Final Takeaways And Farewell

Scott Allen 38:25

Wisdom for Leaders, the official podcast of the International Leadership Association. Be sure to leave a rating, share this episode with someone in your network, and visit the show notes for links, references, and additional resources. Even also subscribing to the Practical Wisdom for Leaders newsletter. For more insights and resources, the link is in the show notes. We leave retreats for senior leadership teams, conduct engaging workshops and deliver installing keynote for organizations, but industries. Until the next time, keep changing money.