The Neuroscience of Learning in the AI Era

Hello, welcome everyone, happy Friday. Welcome back to another week of your brain at work live. I'm your host, Dr. Amasaro, and I am the Senior Director of Research here at the Nierlead Ship Institute. We are happy to have you back for all of our regulars and for our newcomers. Welcome, we're so excited to have you here with us for the first time today. As always, please open the chat box, set it to everyone, and tell us where you're coming in from today. I'm zooming in from just outside of New York City, and it's beautiful Friday. In today's episode, we're going to explore the future of learning in this AI era through a neuroscience lens, lots of research, starting with what the research is saying is happening when we learn with AI and what the brain really needs to deeply encode information or build new habits. So I see a lot of people from all over the place, keep it coming. Joining me today, you know him well, coined the term, the neural leadership, coined the term neural leadership when he co-founded and I'll lie over two decades ago. He is a professional doctorate for an almost five successful books under his name and a multitude of bylines ranging from the Harvard Business Review to the New York Times and many more. A warm welcome to our co-founder and CEO, Dr. David Rock. Hey David. Thanks Emma. Happy Friday. Good to be here. You and everyone, and I've fun to be doing a session with more research. We haven't done sessions with slides for a long time. I think we noticed just how many people were listening to this. It was in the thousands each podcast. So we thought, well, we won't do slides, but then we just slipped into slides today. We'll see. We'll try and explain everything we can see. So you won't miss out. We're not going to share the slides with the podcast, but for those of you here with us in real time, you're getting an extra bonus. So that's what's happening. So Emma, when did you join us? What year did you start with within a like? Well, it's been almost five years. Five years. Yeah. So whatever that was, 21. 2021, right? I just think about all the learning work that we've done and kind of when that was. I think we did most of it well before that. The Eureka scale we published just after that. I think you adopted us with that. Well, learning was something we got really interested in very early 2010, which was like the third journal we published in 2008, 2008, 2009, 2010 was the third journal. We spent about two years on this really big question of just like, how do you make like other foundational principles for making learning stick? And this, this research project was one of the really big successes that we had, especially early on, a bit like scarf where we were able to find a really clear signal in the noise. We were working with Layla Devatch at the time who was then at the NYU memory lab and a couple of other researchers. And we basically said, look, you know, are there a set of key principles for making learning stick? And we found there were some of you will have been playing with the ages model. We'll dig into it. Then we did an update on that in 2015 that brought in some really, really important research on social learning and also insight learning. And then we did some work on like the capacity, brain capacity, something called the fact model that we don't talk about much, but it's actually really important. Coherence, we published a paper on how coherence works, why it matters what it is. And then three really significant papers on insight itself and how insight is central for learning both, you know, where they come from, how to have more of them, why they matter and how to measure them and just like really understanding the mechanics and, you know, learning without insight is really not learning. You need some, you need, even if it's not full insight, not a bigger harm moment, but there's new connections making it being made. It's something called generation. So it's, it's, these are like the kind of five, seven most relevant papers, but we've done lots and lots of other research around, you know, this concept of, you know, how do you define learning, how do you measure learning, how do you, you know, what habits, you know, different intensities of learning, all this kind of stuff. I think one of my, one of my sort of clearest frameworks or most helpful frameworks for me anyway was this concept of easy recall under pressure that I realized there was a continuum from like sort of recognizing something in a multiple choice, you know, question like, oh yeah, that's the answer, right? Recognizing versus actually recalling without help and then recalling under pressure, which is harder and then easily recalling under pressure, which is really the hardest. And a lot of learning really needs to be easily recallable under pressure. If it's, if it's like a way of doing something, if it's a way of communicating, a way of giving feedback, a way of, of running a meeting. So, you know, these kinds of things, if it takes you a minute to try to remember and you only have a member, it's not going to work, right? So a lot of the soft skills that we learn really require that, you know, that kind of framework. And so, you know, learning is such an interesting thing because it's because it hasn't had a lot of measurement around it and because in the sort of wider world, it's something that doesn't have a lot of concrete kind of metrics, companies have, you know, have a long history of just like, oh, well, let's take three days and put it in two. Let's take two days and put it in one. Actually, let's do, let's do a half day online and see what happens and no way of measuring and really knowing if any of these things are working. So we've been trying to address all of that. It's a long opening, but that's kind of where we've, where we've been. And, and maybe I'll turn to you as we sort of start to bring AI into this equation and people start to use AI to, you know, theory accelerate learning or deep and learning. What are some of the things that we're seeing? I guess that's the cliff. Thought is it's making learning worse, but it doesn't happen. Right. Yeah, I think anyone who's kind of kept up with the research has just seen paper after paper come out with how AI is kind of interrupting or disrupting the learning process. So like really easily generalized using AI in this case to a set of participants learning how to use AI to code over time when they used AI, AI assistance, they just learned less when they were tested later. Whereas anyone who learned without AI showed a huge improvement in learning. So it's pretty clear something about the learning process is affected when AI is assisting them. Right. And that's learning to code, but I think that this has been studied in some other domains as well, right? So they're just like learning things where AI just, you know, kind of helps you along, is helping. And what's, what about experience? What did we see there? Oh, yeah. So interesting about this is they did have a huge group of participants and they all came with different levels of experience, different years of experience coding. And what they found is if you, if you took someone with seven plus years of experience coding before participating in this and allowed them to use AI as an, as an assistant, they, they learned less. They learned those coding, the languages, like all of that, they learned less over time than if they didn't have. But it's almost like the more experience you have, the more disruptive AI can be, which is pretty fascinating and scary at the same time. Yeah, no, that's interesting, isn't it? I wonder if it's just they, they're even more keen to be lazy or something. Maybe. They've got a whole lot of experience. This is a really a song-breeding study. Take a take us through this one. Yeah, this was really interesting and just came out. They tested individuals on two different kinds of tasks. So math, like fraction problems and also reading comprehension. This one shows set of individuals going through fractions and they allowed a half of them to choose to be assisted by AI over the course of this. And then without telling them, they took away the AI assistant. And you can see the, the blue line are the controls. They're just learning over time. They get better. They get worse. They learn. So they're learning as they're going through this. The AI assistant, people, people, you could just see they're performing great because AI is giving them the answer essentially. As soon as you take that away, their performance immediately plummets. But what's also interesting about this is they, what's not shown here is they also tested or they also looked at how many problems this group skipped after being losing the AI assistant and people that were used to that AI assistant, they just started skipping problems. They didn't even attempt them. So it's almost like they lost that kind of drive to even try to learn those problems without being assisted. That's the will to think. Yeah. Interesting. So it's affecting people's, like if you, if you, if you, and it's not necessarily that AI is just like giving you the answer, it's like helping you with the process. And if you're doing that wrong, you'll get really lazy and you just kind of won't think afterwards. That's really interesting and important, isn't it? So yeah, let's, let's, let's kind of, we want to set the stage with that and then think about what's actually going on and how's this working. And we, we, we, you know, the first thing to realize is that that learning is a brain thing. It doesn't happen like, you know, on a screen, you know, in your AI tool, it doesn't happen in, you know, the cloud. It doesn't happen on a stage if you're in a, you know, in a classroom. It happens in your brain, right? And it's, it's really interesting. And I, I, I had this inside years, some years ago, I went to a, a chief learning officer conference and I was really excited to kind of, you know, be there and see like best in class learning. It was put on by a reputable company and you know I was excited to like learn, be in the audience and learn how to do learning. I figured out a learning conference, they'd do some best in class learning. But literally for two whole days, the lights were down in the room. You couldn't really talk to the person next to you at all and was just a show on stage, one thing after the other. It was a great show, but like two hours in, I was like my brain was full, there was no digestion, there was no interaction with people, no application, no implications, just like a show. And that can be great sometimes, but it's not great for learning. And so the experience isn't on the stage, the experience isn't on your screen, the experience is in your brain when learning works and that's something we kind of took to heart and thought about. And then this is a piece, some of you may be saying it was a big piece, we published in HBR, I guess a while back now, but it's, you know, what's lost when we work with AI? This kind of summarizes a lot of what we're going to talk about today, we'll put this in the chat as well, but in particular like you're just not getting the activation of your brain. That's the cliff note is your brain is just not activating the way it's for learning when you're using AI. So how does this all work? This is the framework we published in 2010, the Aegis model, it stood the test of time, 26 years it seems, no, that's 16 years, I must be using AI. So 16 years, it's stood the test of time and it's describing the four conditions that you need a lot of to get to easy recall under pressure. So if you just need like recognition of something, you know, from multiple choice, you don't necessarily need like intense hippocampal activation during the task, during the encoding activity. So let me slow down, explain it. If you do want like easy recall under pressure, you need really intense activation of the hippocampus, the network connected to the amygdala, this network doesn't store information, but it activates according to the strength of the network across the whole brain. And so what we know is to remember something easily under pressure, lots of hippocampal activation, lots of networks right across the brain, lighting up when you're learning. And what we found is that you need all four of these conditions, at least moderate and preferably high to get that. So if any of them are low, like if you don't really have emotions, you don't really get this benefit. If you're not really feeling anything, you're not going to get easy recall under pressure of what you're learning. So you need all four of these to be at least moderate and preferably high to actually get this, you know, this domain. And so what we thought we do is kind of take this apart from the perspective of applying AI to learning, right? So kind of what happens to ages when you start applying AI? Talk to us a little bit about attention first of all, Emma. Yeah, so what I love about this model in general is it kind of gives you a checklist, right? Or what are the different conditions you can create? And if we think about AI, you know, and we see the results that AI leads to, you know, poor learning if used in a certain way, you know, what are the domains that that AI is really pulling from and taking away from? So attention is one of the things that we have a super limited resource of, but we also use it increasingly at work. And so it's limited. We're also incredibly poor at multitasking as good as we think we are. We're actually really bad at it. And as good as it feels to do those things, you're not actually going to learn over time. So we're bad at multitasking. We've evolved to be super easily distracted. And we need variety. So as you're building learning, right, we keep keep the session short, we allow to have novelty. We know that people tend to multitask. We reduce that. All of those things more easily engages the hippocampus, right? So keeping that in mind, how is AI taken this way? Yeah, yeah. How is that AI is making this worse? It's sort of it's not obvious at first how using AI would reduce attention. What's the mechanism here? What's going on? Talk us through it. Yeah. So what some of these studies have been coming out recently showing that when people are using AI, they're just getting to this point of what we're calling brain-free. They're just cognitively depleted. And when you look at how they're using AI, it seems that now they're using AI almost as a multitasker. So they're working on an AI tool for some task. And while it's running, they're switching to something else and then they're switching to something else. And then all of a sudden, the work is kind of creeping into times and they would have taken a break, which is incredibly important for your brain to take those short maybe 10 minute breaks or 20 minute breaks. They're getting so much done. So they're very productive, which probably feels very satisfying. But at the end of the day, they haven't actually given parts of the brain that need the break, a break. And they're incredibly fatigued. And even they're seeing people are seeing an increase in employees that are intending to quit their jobs because they're just so burnt out from all of it. And what we think is happening is that it's really the multitasking. That's that's increasing because you're able to now do multiple things at the same time. So you're actually switching your your attentional focus. And that's actually working against what your brain needs. Right. It's it's I mean, hearing stories of people, you know, trying to manage like 10 10 agents doing 10 different tasks and and like switching between them. And each time you actually switch, you've got to, you know, bring all the information to mind and you know, shift context. It you start to get kind of nauseous and overwhelmed. So so it's it's there's there's an interesting effect though. And this kind of explains it. And that's this concept of midterm working memory. If you think of this, there's there's a few types of them, you know, you know, short term memory, which is, you know, the old expression is in one brain and you know, in one ear out the other, right? That was ironic. In one ear out the other is short term memory. So you're like watching TV. You're not particularly engaged. You don't remember what you watched all that well, you know, later on, long term memory is if you imagine this is just, you know, you can see a stage on screen here, right? Feel those of you with us in real time. You imagine long term memories like the audience, working memories like what's on the stage for a short period of time, maybe a few seconds, maybe 30 seconds, maybe at most. So working memory is like, you know, doing math, adding two digits up, you know, you're holding something in your attention. That's working memory. But what happens is when you shift gears and say, right now I'm going to multiply, you know, four digits, the information relating to adding the digits actually stays just on the side of the stage. That's a metaphor, obviously, but it stays in what's technically known as midterm working memory. It's a real thing. And what happens is after you've kind of switched five or six times, your midterm working memory is very full and it's starting to kind of crowd the working memory, the stage itself. As you keep having things popping up. So we've all experienced this, especially at the start of the pandemic, when we started doing like 15 Zoom calls a day and we were just completely overwhelmed by the sort of third or fourth call of 30 minutes and trying to switch, right? So this is sort of back to that point of AI is just putting way too much in midterm working memory on the sort of side of the stage and it's it's reducing your capacity to focus attention on any one thing. And what we see is that really focused attention is necessary for learning. And I'll give you a sort of example of this. For a long, long time, we used to run training programs virtually without a platform without a webinar, we would do it on a phone bridge. And we actually held out for years and years and years after other companies that sort of started using, you know, all the bells and whistles of slides and stuff. So we'd have a virtual group and what we would do is have everyone have to have a pen and paper in front of them. And people in the old days where this worked, people not muted. So they were unmuted. So they couldn't make noise. They had to really focus and listen and constantly call on people like every couple of minutes someone else would be called on. And because they weren't muted, they'd interact quickly to keep the group size small like 20 or so. But between 12 and 20 people, you could facilitate a conversation, let's say a role play, right, a demonstration of something. And people would be really focused. Their attention would be really focused, especially if they know any minute, they could be called on. And as facilitator, you know, move things around. And what happens is because you're not seeing something out there in front of you, you're seeing something really intensely in your mind's eye. And if you facilitate well, people are activating really strong circuits in their imagination, just like when you read a book, right, when you read a book, you don't see squiggles, you see characters interacting. And so as interesting, you know, like you get really focused attention with a single stream of data, just audio, more than sort of diffuse attention across lots of domains. But essentially that's what you need. You're very, very, very close attention, you know, like you pay when you're talking to one person and really focused on what they're saying, that's the kind of attention you need for memory to stick. So talk to us about a generation. Yeah. So generation is the process that your brain goes through to create its own connections. And we do it in several ways, but it is one of the best ways to learn. It, there are many ways we can add to it. So if something is socially relevant, if it's a social context or personally relevant, we just, we create those connections ourselves. We engage more of the brain. Insights are one of the types of generation we talk about all the time because you're actually free. new connections, but learning as an outcome is going to be stronger, the memory is going to last longer. All of those things, you're going to have dopamine rush, you're going to be motivated to drive change. All of those things happen when you have insights. And what's interesting about insights is that in all kinds of generation, is that AI seems to specifically target using AI in certain ways, seems to target our ability to generate our own connections because it kind of does all that for us. Yeah, this is this is the one with the biggest problem. And this is the scariest study I've seen about AI. I'm not a fear monger. I'm overly, you know, overall kind of optimistic about about the potentialities of this tool, it when used intentionally, but we're not using intentionally as all sorts of problems. So this study, if you haven't seen it, it's a fascinating study, 2025. Basically, people were asked to write an essay and then quote something from it. So they're literally, you know, like a exam or, you know, writing a proposal or imagine this is the same as writing an email about a strategy to a colleague. So people are writing something and then they have to quote from it. So the people who just use their brain, 11% of them, struggled to quote. So one in nine people couldn't quote anything. Right. What's interesting is when they use the search engine, that number was the same. So using a search engine didn't cause problems anymore than just using no technology. But when they use AI, 83% couldn't recall anything from what they wrote. Right. And the way, you know, way we understand that is there's just no, there's no like connection happening here. Right. There's no like people are not doing the work to build connections. So that's, it's a really big, it's a really big issue. So the other problem with this and I mentioned a minute ago, you know, insight is the strongest form of generation. Insight is when, you know, a new piece of information comes along from in the world or from in your brain and everything changes. Right. Your circuitry changes, not everything, but how you understand that issue changes in a meaningful, meaningful way. It releases a lot of dopamine and neuroprinephanephan and stays with you, insights stay with you, sometimes forever. So when, when AI is solving for you, you're not having insight. And that's really a problem because insight, you know, is the foundation of getting smarter. It's the foundation of engagement and motivation as well. Most people their whole career started with an insight they had. So it's a very, very important thing that we don't rob people of insight using AI, but that's very much what's happening. And you want to talk through this study. This is a crazy one. Yeah. It's also a great study. Two groups of people asked to invent a new toy with a couple of objects. You could use chat GPT or not. When you didn't use chat GPT, everyone's ideas were different, ideas were different because they're all creating their own ideas from all of their own experiences without any support. But when they use chat GPT, because chat GPT is summarizing all of the ideas in the world, 94% of the ideas overlapped with others and what's really even more interesting is nine of them actually independently gave their toy the same name. So they just copied each other without realizing it because all they were using was actually just creating a sum of everything that was out there. Right. So people are not having their own ideas. They're not having insights. And they're taking the sum of the average of all information and then doing that. So that's interesting. Now this is the scariest study of all. And we just, I think this is coming out in fast company shortly. If you follow us on LinkedIn, the company or me, NLI or me, you'll see the pieces as they come out. We've been publishing a lot. And I will say we're not publishing more because AI is helping us write. We don't use AI to write. There's just so much happening in the world. And there's so many interesting studies coming out that we kind of accelerate. But this one has been like blowing our minds for a while. So it's a study published in science in just been March this year. And science is an incredibly reputable publication. And it basically explained like step-by-step what happens when people use mainstream AI's, which are also panicked by nature. When people use mainstream AI's for interpersonal issues. Right. And the business model of the big LLMs, whether it's Claude or ChatGbT or any of them, is to keep you engaged. And one way it does that is just agreeing with you. Now that's problematic in some realms. But if you're using these tools for soft skills, right, or for leadership development, right, or if you've got a leadership question, right, you're going to these tools. What will happen is you're always right, which makes everyone else always wrong. So you're not going to learn. You're not, you're like, you're going to feel better, but you're not going to learn. So a sick of phantoms tool that always agrees with you is going to actually reduce learning. And it's going to probably increase the likelihood of toxic leaders and toxic cultures, because everyone's going to just blame everyone else. So you're not getting that like, you know, rupture and repair and insight and learning and coming back together. So it's really a it's a, you know, it's a really big big challenge there. Anything you want to sort of say about generation before we go into emotion? Yeah, other than, you know, it's a great way to design experiences to learn. And we'll actually talk about what's interesting about generation is there many ways to use AI to encourage your own generation. So this is where it comes back to how are you using AI? And because I know a comment in the chat about the different kinds of AI users. And there are some that use it in a way as a partner. And that's where you can really still have your own generation. So it's really how you how you're thoughtfully using AI here. Yeah, but really intentional about it. We're definitely going to give you some tips on how to think about that. And we're kind of studying the problem first. Then we'll get to some solutions. But they get support to kind of understand each of these elements. You know, attention is being really dispersed to fuse, you know, not focused. Generation is not getting activated on this really intentional. And emotion is not really happening. And that's that's a problem. Emotion is essential for learning. So I recently decided to get a finite license to do some things with with my own funds. And I realized I had to go through an exam. And I hadn't done an exam since like decades, decades of like, oh my god, can I still study? I've like, I've got to find the fastest most efficient way to do this. And I was thinking about ages and I realized that emotions are really, really important. If you don't feel emotions when you're learning something, you just don't learn. I thought, well, the best way to to have emotions is to test myself when I know nothing. Because every time I get the answer wrong, I'll be really upset at myself. But every time I get the answer accidentally right, I'll be really excited. And then when I come back and in a long time, I studied for these exams, I didn't read a single thing. I didn't like do any normal studying. All I did completely was test and test and retest and retest. And the reason I found that worked is it just it activated so many emotions. It was like, you know, I was really observing myself. And it now, interesting point, I used AI to build these tests and to get them refined and that saved me some time. But I was doing the tests, right. And we'll come back to we'll come back to this. So just, you know, understanding, understanding emotions and just how important emotions are was, was, you know, really, really helpful. Do you want to talk through a little bit of the science of the summer? Yeah. What's interesting about this is that, so we're always thinking about the hippocampus when we think about learning. What is engaging the hippocampus the most? And the amygdala, which is what the area that kind of engages when we are emotional, all of the text emotional context is right next to the hippocampus. So if we're engaging it, whether it's something positive or something slightly negative, like testing yourself and getting questions wrong, that like angst that comes with that, both of those things are going to just kind of more greatly engage the hippocampus. So any kind of kind of emotional environment could be something that's slightly happy, slightly like nerve racking. That's why sometimes testing kind of encourages someone to learn. So it can be both positive or negative. That's just going to more greatly engage with hippocampus. And we know that negative emotions are stronger than positive. So like in this virtual classroom, I was talking about earlier where you're only just listening to things. The fact that you can't see people and the fact that you could be called on actually made people really anxious. And in more recent years, companies like Wooden Letters do it anymore because people explain too much, complain too much that they were too anxious, right? I don't know what happened to the world. But the maybe as opposed pandemic thing, they just weren't, you know, the point was to make people uncomfortable, but just a little bit because a little bit of comfort while you're learning something activates the hippocampus in a really good way, you only need a little bit of discomfort and you get there, right? Well, you need a lot of positive emotions. So one way or another, you've got to get some strong emotion. And we're seeing that using AI to learn has some negative impacts on emotion, increasing loneliness, boredom, you know, reducing enjoyment. Talk us through this one, Emma. Yeah, this one, this study actually looked at the drop in like a rousal, an increase in boredom. So when you're using or collaborating, in Gen AI, you're going to see this immediate boost. We see this across the board, right? You're immediately performing better. But then if you take away AI, which we saw in those earlier studies, you lose that help. You don't really learn the skill. And then over time, if you continue to use AI for tasks, people start reporting that they're not motivated, they're bored. Their arousal level is just dropped. Because you're not getting that stimulation that dopaminergic rush from actually learning the tasks that you're doing and succeeding at them, having that purpose associated with the task, so you're just bored. And losing that, that's part of the emotional aspect. Yeah, bored is not a good way to learn. I remember trying to actually, when I was studying at this exam, I remember trying to read some summaries and stuff. And it just literally put me to sleep. It was like, I can't do this. I have to-- and I was so clear nothing was sticking. So boredom is not good for learning. And it's also even reducing the enjoyment of the experience as well. Talk us through that. Yeah, this was that same study I showed earlier with the coding experience. They also asked them later, how much did you enjoy doing this with AI or without AI? And people reported consistently less that they actually enjoyed it. Even though they had to help and they were probably getting the tasks faster, they just didn't enjoy it as much. Yeah, yeah. So the emotions are not great. Now, what about spacing? So spacing is such an interesting one. One of the interesting confounds with spacing is people predict that learning in a block, just before you need to know something, it will be better. And they're actually right. If you just need to study for an exam, it's great to do it just before. But you won't have long-term memory. So you won't remember a month later. And I might have hacked this one as well when I was studying that I did learn in a block because I know that's better. But I probably can't remember a lot of things now. It's just not something good to admit, I guess. But with spacing, you really need that coming back. And there's a number of reasons that that actually helps. It's based on the fact that you don't store memories. You actually grow them through attention. And coming back to a memory in a different context and re-growing it, re-watering it is really powerful. So what's AI doing for spacing? Yeah, well, we talked about this earlier with this intensification of work. It's now you're using AI even during your break. Work is just kind of seeping into all of those natural breaks. Like maybe you're checking on the project that you're working on with your agent when you're supposed to be in between meetings, let's say. So it's just taking away those natural pauses. And that is taking away the time that your brain needs to consolidate that learning or to kind of like turn your attention off. There's so many reasons why we need that spacing. Yeah, it's interesting. And I want to comment on something in the chat people are talking about, chat scripts and stuff. The answer to this is not just like getting AI to challenge you because it's still fundamentally is sycophantic. It's very, very hard to get the sycophantic. You can turn it down a bit, but it's very hard to do that. And there are ways to use it that will deepen learning. And we're saying intentional use. So this is not a fader-complete that AI is going to be terrible for learning. There are ways you can actually make learning better using these tools. But it's much more than just tuning your tool of choice. There's some kind of bigger perspectives here. So let's dig into that. Because it can increase ages and amplify learning. Let's talk about this high level, first of all. There was a hinge in this study. Exciting. Yeah. So what this study did-- so this was the study where they were either doing reading comprehension or algebra. And they found that over time, the people that had the AI assistance, they kind of dependent on it. Well, they pulled this group apart based on whether they asked for the answer or whether they asked for hints or clarification. And when they pulled that group apart, they found that those individuals that just asked for a solution, they just went right to the answer, learned the least and skipped the most. And those people that just asked for hints-- or let's just-- can you clarify a bit-- they learned a bit more, not as much as those people that didn't use AI at all. But you could see that when you're just using AI for a bit of support, clarification, some hints, like help me along the way, they did seem to learn a little bit over time. Right. So as scripts, like just asking your tool to be less stochophantic won't do much. But if you said to your tool, don't give me the answer, give me some clues. But explain what's going on here more so that I could come to my own answer. Yeah. Can you imagine? Yeah, this is kind of getting at that generation aspect, too. You're actually encouraging generation. You're asking for some questions as opposed to the answer. And so you're forming your own answer a bit. Yeah, and it's a really important one. And this is what I did when I was studying for my exam is that as I set up questions in Gemini, I set up multiple choice questions in Gemini. And if I wasn't getting it, I had the option of a hint that would kind of get me partly there. And then I might still get there or I might not. But I'm making the connections. And it's such an important phenomena. Am I being given the answer or am I coming to it through making connections in my own brain? It's such a central mechanism. And you're going to explain this one. Yeah, it's the same kind of thing. And this is a totally different study. Showing again, we can lose our ability to generate our own answers. Here, two groups. One group were asked to make decisions. It was a health-related decisions based on an AI explanation or an AI recommendation explanation. So the AI gave them, like here, we recommend you do this thing. This is probably the best decision. Or here is all the research into it. You can make your own decision. And then after the fact, they basically tested their learning on all of those that information and the decisions they made. The group that was given the recommendation to make a decision didn't learn as much. You can see that overall learning is less than if they gave them all of the information, but didn't make the decision for them. So taking that last step, generating your own decision or your own learning from the information still kind of allows you to maintain your ability to learn from that information. Yeah. So a recommendation is kind of an answer, whereas an explanation is like, look, here's what's going on. See if you can come up with it. And so AI can be really effective at doing that. So let's kind of summarize what intentional AI use would be for learning or what we've been calling human first AI adoption for learning. So what does it look like? So you're using AI tools, firstly, to increase focus attention. So this is really important, right? How do you use various tools so that you're actually focused more? And that's an interesting challenge, because you don't want to be multitasking. You don't want to be between things. You want to be using AI to focus much more. Secondly, you want to be using it to deepen the insights you have. So the insights are stronger. You want to be using it to activate emotions. You can imagine like role-playing things with an AI, it can really deepen emotions. Like playing different-- practicing different things. And also to enable spacing, like AI can come back and remind you of different things, bring things back to mind. So you want to use AI tools to actually enhance AI tools, basically, because AI isn't changing. The human brain isn't changing when it comes to learning. In fact, the human brain hasn't changed for many, many thousands of years, many, many, many, many, at all. It's pretty consistent. So what the brain needs to learn is going away. And we need to kind of anchor these tools on this framework and use it to increase attention, deepen insight. And you can hear the insight one in the last couple of studies, like you're making the connection. But also giving an explanation and not giving the answer also increases your attention, like you're focusing more. And it's activating emotions and all that. So it's an important insight. And a couple of ways to think about this, increased attention will come from really personalizing any kind of learning. So I'm speaking to folks who are learning designers, learning architects, all of this. The more relevant it feels to a person, the more attention they pay. The more it's relevant to them and personalize. It's incredibly different when something feels like an issue that you're dealing with this afternoon or tomorrow. OK, my attention's really here. And all the other elements of ages, right? Safe practice, fantastic for insights. Real time feedback through AI tools is very powerful for emotions. Hey, let's debrief that call. Let's debrief that meeting. Let's debrief that one-on-one with your AI coach. These things will activate really strong emotions if they're done well. And then, generally, the strategic use of AI coaching will enhance spacing because instead of having just one hour coaching session every week or two weeks a month, you can have quick AI conversations, quick AI coaching conversations, quite a few times a day to practice things, to deepen learning. So the thoughtful use of AI coaching can really help with this. And what's interesting is that the strategic use of AI coaching can do all four things on the screen. the right here. It can personalize it more, give your chance to practice, give you real time feedback. And that's going to increase all of ages. But there's a sort of dark side to this. It doesn't happen just by giving everyone an AI coach and nudging them and encouraging them and all that. There's something else that's necessary that we'll get to. But any comments on this before we go on, I think this is sort of a central point. It's about ages, you know, it's about maximizing ages through. Yeah, I mean, I think that it's I just like we would talk to people before before AI became so evident in everything that we do, how you're designing the learning, you can specifically design it to engage the hippocampus in the way that like ages kind of speaks to. So in all of these ways, AI can be fabulous for for learning, but it really does take this like intentional planning before beforehand. Yeah, so thinking it through. So and let's let's kind of unpack a couple of really must-do's. So the first one is, you know, make any kind of learning incredibly relevant to the person like super relevant to things they're doing today, this week. What that does increases attention, generation, animation. I keep mentioning the testing effect. It's such a it's such a valuable. It's a tried and true method for really making learning stick leverages attention, generation and emotions in a big way. Allow for insight through reflection. So this is a big one. So you're designing learning with AI tools. You want to make sure there's there's reflection built into the the the the the the the process, right? It might be, you know, reflection questions, people have to answer overnight conversations. They have to have with other people, but really making sure reflection is built into the process. That's going to increase that's going to increase generation emotion and spacing and for those who miss the first half attention generation emotion spacing is the ages model that we're kind of referring here by letter. And then a really, really big one here is keep a human in the loop. And this gets the positive jackpot of ages that when a human's involved, you're paying much more attention. You're generating connections through speaking out loud. Your emotions are much higher and you're getting this interesting spacing effect number of ways as well. So so we actually want to say if you're designing learning and you want to incorporate AI, don't take the human out of the process. Have humans in the process strategically placed. It's a it's a really, really important thing. Actually, we should have included included this study, but there was something just came out in the last month that I'm going to paraphrase because I read it closely, but it was basically saying that the way companies are using AI coaches and individuals are using AI coaches, they get a little like they'll solve a problem, but won't do anything about it. They'll have an insight, but not take the action. And you know, the situation didn't change. And from my perspective, that's because they didn't have a human involved because the number one driver of human behavior is not wanting to look bad in front of other people, not wanting to either be different to others, which is relatedness and scarf, or not wanting to feel worse than others, status and scarf. But it's these social emotions that occur when a real human's involved. It's so, so important. And so designing with AI, keep the human involved. Now, so, you know, we're a big fan of AI coaches. We have an incredible one. If you haven't played with it yet, get on there and do it. Niles, our neuro intelligent leadership enhancing system, Niles is incredibly powerful. 80% of conversations are resulting in insights. But if you don't, you know, if you don't think of Niles as part of a system that includes humans, you're probably not going to get the results you should. And if you don't think of it strategically, like any AI tool, you might not get the usage. So you want to really think about very strategically. So by the way, I will, we'll, we just put a link in the chat. I decided to give everyone a 30 day free access to it, rather than a week, just so people can really get to experience it and play with it and try it. So if you go to askniles.ai, ask, so ASK NILES askniles.ai, you'll, you'll be able to get a free month to play with Niles. But, you know, whatever your, whatever tool you're using, you actually want a whole learning ecosystem because a tool like Niles can do these things. But you really need to keep the humans in the process. And let me explain, explain why. And this is in fact the second paper we wrote on ages. It's from this second paper. There's a whole lot of things happen. The first one, and I wrote about this in the HBR piece that when you speak out loud to another human that, you know, at least pretends like they care, you, you, you get like this thing called spreading activation to the networks are stronger and therefore they connect to other networks in your brain. Right. So you, you're, like thinking a thought versus speaking that same thought is very different. It's not just the speaking lasts, you know, 10 seconds thinking lasts a 10th of a second. That's part of it. But it's much more intense to actually speak a lot more of the brain gets activated. Remember the hippocampus and the strength of the network right. So speaking to another human just creates much more of a network. And then you get what's called this spreading activation benefit of it connects to other things. So when there's no human involved, you're not really getting that same thing. You'll get some of that benefit from from an AI coach, but not, not all of it. Secondly, very interesting thing when you believe you're talking to a human, you, you, you store information in a completely different brain. Not a different brain, different part of your brain, but it's a different memory structure. And you can remember like, you know, how that person was feeling. If you saw them, you know, what they were wearing, what their mood was like, you know, like you remember all this information automatically, it encodes automatically. It's and it's connected to that network. You're not really going to get that with an AI coach. You're not going to have all this social information. So and the third one is probably the most important one is when there's a human involved, you're getting what's called positive social pressure, especially if it's a human you don't want to look bad in front of for like an important person in the organization. You've got to do things because you don't want to look bad. And what we're seeing is you don't get that effect with the AI coach. There's no like social pressure to do your homework, basically. And so we think you need to keep the human involved. And would you want to add anything to this? I mean, you know, jump into some of this, but yeah, I mean, this is, you're speaking to all of this, but I just one thing to kind of keep in mind is that we're social creatures. We've evolved to survive in social groups. So so much of our brain is social. So if you add a bit of social to any experience, like you're bound to remember it so much more, especially if it's social and people that are relevant to you. So your leader, your team, if you're, if there's a way to involve them, especially someone that you respect or someone that you don't want to look bad in front of, that that account social accountability is huge for drive learning and then getting action from it. Yeah, so you really want to keep that in the network. So so this is this is now going to lift up a bit. And we want to give you a bit of a provocation about learning in the AI era. Okay, coming circling right back to where we started, like what about learning now? So right now your learning function in your organization is focuses on, you know, selecting and managing vendors, project managing, participation in in those programs, whether they're in person virtual, whatever, but you know, you're basically project managing. And then you know, tracking and reporting and tweaking and all of this. And it's that's that's like where we are at the moment in learning. But we think there's a big shift to really understanding the science of learning, identifying key audiences and times. Now what do I mean by that? Like, you know, you know, there's a lot of things people could learn, but the things that, what are the things that really matter? Right? If we're talking about retail staff, they need to know how to sell your product. They need to know how to deal with customers. Like, you've got a lot of turnover of retail staff. If you've got a lot of stores, that's a key audience. That's a key time. You know, the first month of a new employee, you know, really teaching them how to, you know, interact with customers just right. You could make a big difference, you know, at that point or a first time manager or, you know, or it could be anyone onboarding into your company. But identify the key audiences, the key kind of points in time where they really could benefit and then build these integrated habit activation strategies that integrate both an AI coaching tool and a and humans. And I'll give you some example of how to do this. But I just want to say and that that I think in the next, you know, year or so, there's going to be a lot more media, a lot more conversation about leaders not being allowed to use, leaders and managers not being allowed to use the mainstream elevators for interpersonal issues, because it's going to be a problem. So you do want a specialist tool of some sort. You want a specialist tool that it's not just isn't sycophantic, but actually calls people out, challenges them. You actually want a tool that helps people learn and challenges them. And that's just really hard to weave into a tool that's that's fundamentally trying to keep you happy. Anyway, this is where we think learning may go. And the internal teams building these whole architectures and leveraging various specialist tools to accelerate learning. And so there's a shift from just in case learning design. So training programs, in okay, you need them. Quite general programs, just having to continually nudge people, not much integration. What we see from learning research is about 80% of the learning doesn't stick. So you're not getting easy recall under pressure with the majority of it. People forget 80% of what they're learning, which is not great. Whereas if you can get to real time learning design, right? So using an AI tool to like in real time, teach someone how to sell, real time, teach someone how to code, develop these highly personalized sprints. So what you'll get is much more compelling participation and you'll be able to integrate into work. So this is about like an AI tool that in real time can challenge people, real time can stretch people. And what you should see is actually 80% of learning sticking. So it's the opposite 80% vanishing versus 80% actually sticking. So and what's happening on the right-hand side here is very, very high ages when you design this right. And I'll give you kind of a big idea we've been playing with. You can take any content area, could be technical skills, could be any content, but it needs to be for a particular audience at a particular point in time. And I've got some examples on screen here. Could be on boarding, could be sales training, but it's a specific audience at a specific point in time. You can work out what we're calling a sprint, either a 30 day or a 90 day sprint, that leverages an AI coach, that makes it very personal, but still has the human involved. And the concept that we have for this, we're calling the press framework or press strategy, you're actually, you're not just giving a tool to someone and saying, hey, you know, go and improve your sales skills, you're actually designing an experience of 30 or 90 days. And that experience involves custom prompts that you're getting people to use. So you're giving people a series of prompts. They can choose from a couple, so they have some autonomy. But, you know, week one, you know, if it's an AI, fluency sprint, week one's gonna be, you know, have a conversation with your AI tool about where you are right now, and what you're strong in around using AI, and you know, where you think you need work. So, so these prompts will be designed over that month for three months. And then you'll also have these required activities. For example, obviously if it's an AI fluency sprint, it'll be, you know, use one new AI tool you haven't used before. Now, we'll be thinking about how you create this positive social pressure. It's a concept we call sendbacks to term we call. So a sendback is a summary from a conversation. It's not a whole conversation with your AI, but it's a summary of a conversation, like maybe at the end of a week or two weeks or a month, a summary at the end of a period that summarizes your learning, that then goes to the sponsor, right? And that sponsor should be someone who has some coaching skills who's, you know, good at interacting. So, so this architecture, we're using this to design these 30 or 90 day sprints. And, you know, as an example of 30 day sprint, can radically deepen AI fluency, right? You designed for beginner, intermediate and advanced, and you can get this incredible impact on AI fluency. Using an AI coach, little bits of content dropped in as well. So just a little bit of content, for example, about, you know, partnering versus offloading. It's not teaching them to, it's not teaching people to prompt. It's getting them going, it's getting them moving. So this is an architecture for getting learning happening. So a little conceptual right now, if there's anyone interested in kind of really digging into this, we can show you in real time how to apply this to any learning content, like onboarding training, in any kind of training, including AI fluency. Just put the word sprint and your company name in the chat. Someone will follow up with you afterwards, and make some time with you. We'll walk you through a specific example and show you exactly how you can design this. But the big idea here is over time, challenging people to focus, getting people to focus in just the right way that they need, but also all at once with others. So this is, you know, this is, I think, where learning's gonna go. You know, the AI coach is including us, you know, some of them are amazing, where, of course, biased, but Niles literally, you know, eight out of 10 interactions that provide breakthroughs, it's pretty amazing. But just putting an AI coach in isn't really the answer to transforming learning. To transform learning, you want to design these architectures. We're calling them sprints. You want to design these architectures that keep the human involved and keep, you know, exactly the right kinds of prompts. So yeah, see some folks interested in that. We'll follow up with you on that. Let's just wrap this up. I know we've got some announcements as well, but what we're imagining, and, you know, it's early days, but what we're imagining is the whole field of learning, in particular, like your learning organization in your company is gonna change. It's not gonna be as much, it's just not gonna be as much, you know, project management and vendor selections can be much more involved in, like working with people and much more like designing interventions using tools. So you're gonna be empowered with more tools to design interventions and manage those interventions, but really focused on building habits. So that's where we're imagining it going. Any thing you want to say that as we wrap up, we'll go to some examples. - Yeah, I mean, just the only thing is if you go back to the foundation, keeping the brain in mind, you're always going to do what you need to do what you want. And in this case, create lasting learning, right? And lasting memories. So I think if you're designing anything and working with the brain as like your anchor, then you're going to succeed. So I think that's lost sometimes in how others have designed it. - Yeah, I know that's how we're thinking about it. And if you really start with the brain and start with ages, you'll see that the clever use of AI coaches with humans over time, with a little bit of content and just the right prompts, you can actually do some incredible learning, much cheaper, much faster, much better, much more personalized, all of that. So yeah, anyone interested in kind of digging in, put your company name, preferably your company name and the word sprint, we'll one of our folks will come to you and set up some time. And a couple of quick announcements, we just launched an AI masterclass for HR leaders. So any senior HR person could be CHRO or one or two down, but we just launched that. In fact, yesterday we ran the first kind of in-person version of that, but July 1, 8 and 15, if you're an HR executive inside an organization, we've got a three module program to really understand, not just learning, but all of AI transformation from a brain perspective. So that's a new program from us. That, I believe that QR code works or just go to neuroleadership.com. So that's a program we're just launching literally this week, three 90 minute modules virtually with assignments between to really get your head around the brain and AI transformation. And focused on AI fluency for the individual AI transformation for the organization as well. We also have, if you're interested more in design and facilitation, we have another cohort starting of, we've got a BBDF, brain based design and facilitation that's coming up. I think we just put that link in the chat. We're doing another version of the mini summit we ran yesterday in New York. We're doing another version of the Chicago also. I'm about to go on tour and do them in Sydney and Singapore and London. So if you have colleagues in Sydney, Singapore, London, we're hosting these there as well. And then finally, big announcement. Emma, do you wanna tell us about your partner? - Yeah, someone's coming up. We have a date, the 14th and 15th. It's free so you can register already. It's on our website. And it's gonna be two days of all the best from us and the latest science, the biggest thought leaders. Yeah, so we're excited. You can register right now. - Yeah, we have the dates held October 14, but we did something crazy this year, we've made it free. - We have free. - We can get it out. The most possible people. We were getting a thousand or so people, everyone, but we think this is just relevant to a lot more. So you'll be able to log in and share this widely. So thanks everyone for being here. These are big ideas, important ideas. Emma, thanks for the amazing work partnering on all of us. And I think there's a poll to go up. As you still hear, we've got to do it. We can throw the poll up. Let us know how to support you. We'll leave the line open for another minute or so while people do that. Thanks so much, Boggen behind the scenes and Emma, have a great week. All of you, I will join you back in a few weeks after travels, but there's some great sessions coming up. Emma, do you wanna tell us briefly what is going on? - Yeah, well next week is Memorial Day weekend. So we're on another little break. And then the following week we have a session on our high performance GPA, growth mindset, psychological safety accountability. We'll talk about on latest data we're collecting, on leadership now in that sense and how it relates to AI integration. - Amazing, thanks so much. Thanks everyone. We'll leave the line open. We'll leave the platform open so people can finish that poll. Otherwise, we'll see you in a few weeks. Thanks for having me. Bye bye. (upbeat music) [BLANK_AUDIO]