Frederick Wehrle: Optimizing Adult Learning with AI

Frederick Wehrle, Assistant Dean for Academic Innovation and Learning at UCLA Extension, discusses his research on using AI to optimize adult learning. Wehrle focuses on developing AI-powered instructional tools and approaching learning and course design from a neuroscience perspective. He shares practical strategies for educators and learners to keep pace with the changing needs of industry, including advice on the best ways to use AI for learning and development.    Transcript Julian Alssid: Welcome to Work Forces. I'm Julian Alssid. Kaitlin LeMoine: And I'm Kaitlin LeMoine, and we speak with the innovators who shape the future of work and learning.  Julian Alssid: Together, we unpack the complex elements of workforce and career preparation and offer practical solutions that can be scaled and sustained.  Kaitlin LeMoine: Work Forces is supported by Lumina Foundation. Lumina is an independent, private foundation in Indianapolis that is committed to making opportunities for learning beyond high school available to all. Let's dive in.   Julian Alssid: We spend a lot of time in the show, in our consulting engagements, talking about best practices for building effective, engaging programs and learning experiences for adult learners. It's a topic I find endlessly fascinating. Kaitlin LeMoine: As do I Julian, building programs grounded in the principles of andragogy, and allowing adult learners to demonstrate what they know and can do is so critical.  Julian Alssid: Couldn't agree more. And I often think back to our time at Southern New Hampshire University, where we helped to build College for America that project based, competency based online program was a real breakthrough moment in higher ed.  Kaitlin LeMoine: It really was. College for America was designed to reach adult learners who had many skills but had not previously earned a college degree. The program offered them a way to demonstrate mastery of competencies through projects aligned to real world deliverables that met academic requirements in an online and asynchronous format.  Julian Alssid: Our guest today is doing some really fascinating work in the space of adult learning and course design, taking the notion of learning, design and optimization to a whole new level.  Kaitlin LeMoine: He certainly is. We're excited to have Frederic Wehrle on the show with us today. He's the Assistant Dean for Academic Innovation and Learning at UCLA Extension. Frederick joined UCLA in 2023 after serving at UC Berkeley Extension since 2018. Before immigrating to the US, Frederick led accreditation and International Relations at business schools in Paris, France. He's held faculty and administrative positions in France and in the US, and served as an advisor and mentor to startups, nonprofit organizations and universities worldwide. Grounded in research on innate and in learned behavior, Frederick focuses on developing AI powered instructional tools and applications of those tools to adult learning. Essentially, he's approaching learning from the perspective of neuroscience and exploring how we can optimize it. Welcome to Work Forces, Frederick.   Frederick Wehrle: Thank you. Thank you so much for having me, and thank you so much for the very kind intro. I'm very excited to be with you. Julian Alssid: Well, and we're excited to have you here as well, Frederick, welcome to get us rolling here. Caitlin said a bit about your background, but we'd love to hear in your own words about your background and what brought you to your work.   Frederick Wehrle: Most fundamentally, I would say my background is in behavioral sciences, human behavioral sciences. I specialized during my studies, actually, back in Germany, on behavioral ecology, neuroscience, anthropology and bioinformatics, and my key interest there was innate behavioral patterns and pre existing biases. So things that we are born with in terms of mental pathways, and I've been fortunate to be able to do a PhD in Paris at the Sorbonne where I was specializing in consumer behavior, which is kind of this subgroup of management and marketing sciences that actually looks how marketing and marketers are able to manipulate people, and then tries to explain how that works and give the tools to corporations and policy makers, I would say, to regulate if necessary. So when I was saying hey, I actually study, from a biology or biological perspective, how humans react without knowing that they do or act without knowing that they do, that was very interesting for the people in that field. And so that really was, let's say, my education background, I was able and lucky to become a faculty relatively early on in business schools in France, and put my my work to practice and apply a lot of this neuroscience into my teaching, and then very quickly, was asked to use it to design courses, design programs, to design entire degree programs, review the entire structure, if you want, of of schools through accreditation processes, build new types of partnerships and so on internationally. And when I got the opportunity to move to the US in 2018 there was an opportunity at UC Berkeley, and their continuing education division, extension division. I was very excited, because I've been I've been concerned with AI laying off millions of people or displacing millions of people, 10 years ago already, and I was mostly concerned because I didn't find anyone particularly at that time concerned. And I thought, with the background that I have and the theories that or the research I was pursuing, I'm effectively on to figuring out how we can make education at least three times more effective and efficient. So here in California, first at UC Berkeley and now at UCLA, this is really what I'm pursuing, and that is really figuring out how how to teach differently so that learning becomes much, much more effective.   Kaitlin LeMoine: So can you tell us a little bit about what your current research is focused on and what you're learning. I mean, I'm intrigued by the concept of, you know, making education at least three times more effective, right? So, like, let's, let's dive in there.   Frederick Wehrle: Despite the fact that it was over five years in the Bay Area and Silicon Valley and so on, I'm still humble. I'm only 3x ing. It not 10x ing. And yeah, they would not be happy. I couldn't pitch this to Google, probably. But the interesting perspective is that, unfortunately, let's say neuroscientists really haven't looked into how learning happens naturally through most of our lives, before the late 2000s, so about maybe 15 years ago or so, and actually started working on animals, and later figured it out with humans. And it's quite fascinating to see how that relatively recent research paints a very different picture of how we learn. For me that can be easiest understood if we think about what we're doing right now. So we're having a conversation. We're talking about things that we have some mental representation of education. Our brains are active. And in this situation, everything I'm saying, I don't have to repeat. You don't have to write it down and put it on a Quizlet card and learn it by heart and so on. Everything I say, maybe with one or two repetitions, you're able to assimilate. That's actually the proper term for this process in neuroscience, you assimilate that information into an active framework in your brain, a schema, and that active framework adapts to that new information, and so it's memorized. And so effectively, this process of learning, contrary to what still most education scientists and learning scientists believe, does not have any blockers. So the system that most education is still built around is a system of repetition. Because also in neuroscience, we were able to prove that if you have to learn something anew, completely anew, without any context and without any mental frameworks active or schemas active, you have to build those frameworks. You have to repeat, repeat, repeat, and that's how synaptic pathways are built and reinforced until they they're there, and that is usually now, if you look at the broader picture and child development or human development, that's really just how humans learn very early on in their childhood, like the famous light switch that you kids click 100 times until they figured out the principle of the light switch, right? But once you have that framework, once you like this basic pathway is built through repetition, and then, especially as you have kind of finalized building those frameworks, you're really switching most of the time, throughout your life, into this other pathway that has been discovered only more recently, which is information is just assimilated into active frameworks and those frameworks adapt unless You sit at school in a classroom that is sterile or at least out of context, and you learn from a book out of context, from a person out of context. And for me, the these things were, particularly, as I say, egregious, and became very, very clear when I was working in business schools back in France, or early on, even in my career, because in France, like in Germany, you have a system, or systems that are very popular, which is called apprenticeship models. And the most fascinating thing there is, is that you have a very large amount of students studying their entire three year bachelor program (Europe is three years for bachelor's,) or your two year master, either full time studying or in an apprenticeship where you work about 60% of the time and then you study about 40% of the time. What is fascinating about this is, all students finish on time. If I share this oftentimes here in the US, people like pause and scratch their heads and say, how's that possible? Especially like, there are institutions that really are extremely proud and adamant about saying that you really need to be full time otherwise, like you can't dedicate yourself and so on and so forth. And I think that's true. If what you're studying is sciences, where you actually spend a lot of time in sciences or even humanities and so on, but you are, like doing it, and you're preparing for university environment you want to become a researcher, then, yes, you're effectively doing an apprenticeship in in the university, but for many, many other fields, it doesn't make sense. So this is the background, if you want that I'm basing my work on. And I find it particularly fascinating to think about workforce development and to do this in continuing education, because the, let's say the good instructor and the good learning designer in this field already knows that the best and most efficient way to teach is to be able to kind of capture the students where they are, is to be able to understand who they are like with what knowledge do they come in there, and then, if it's in a life interaction likework with with what they bring to the table and their experiences, and organize in a way where they can share and practice with their experiences. So this is a very good background for me to to push that forward and formalize it a little more, to be honest, though, I must say, much more of a experimentalist than I'm a theorist. So I do spend much more of my time, or have over the last decade, experimenting, building the types of courses, running them, seeing how they work and so on, rather than spending my time writing literature reviews and contributing to a relatively set field of research, which is all based on a very different approach to education.  Julian Alssid: So given your your approach as an experimentalist here, and we're, you know, Kaitlin and I are also very grounded and in this kind of bringing together, this intersection of work and learning, which we think you know clearly has to be theoretical in many respects. But ultimately, people need a job. They need to apply themselves. And flashing forward to from eight years ago, when you saw the future, what do you see as some of the greatest opportunities for using generative AI in learning and course design, and also, conversely, what are some of the biggest challenges? Because it seems to be the story of the day.  Frederick Wehrle: Yes, it's the story of all our lives. So I'm actually hopeful that through the work I'm doing, also with colleagues, we're able to to provide some some way through this, I can maybe start with the challenges, because that's maybe top of mind for many people. In my experience, interestingly enough, I think the biggest challenge comes from a lot of resistance from the people that are the faculty, instructor slash expert, or also the instructional designers, the learning designers, who I think, like in many professions, don't feel very easy about the introduction of AI because it seems like it is getting at the core of what they are doing and what they're contributing. And having spent a lot of time, evidently, like overseeing these types of teams and being with a lot of colleagues, I realize it's justified in a certain way, because nobody has given many people that are not deeply into AI a framework and somewhat of a roadmap to where this all will go, and it's certainly not a goal of what this will look like at the end. So if you want as an as an experimentalist, the beauty here is that I was able to work out, and am working out with some colleagues of universities across the US a framework on how to effectively use generative AI for course content development, either in the role of a subject matter expert, so the AI can take the role of subject matter expert, and an instructional designer can actually go and build the entire baseline of a course, if you want, that then gets reviewed and analyzed by a human subject matter expert, or in the other form of an instructional designer, where subject matter expert has the help to effectively design the entire baseline of a course with generative AI. The point of us coming together is to figure out how to do this in a structured way, provide a framework how that looks like, and demonstrate how you get to a very high quality output and how much time that can save you, for example. Now, the interesting thing is, we're doing this because at that point the entire all the professions in the entire industry will say, like, okay, so we can actually get this if we do XYZ, if we follow that pathway. Now we can adopt it without having that much fear, because we see that all the jobs across the entire if you want production line are maintained. And the beauty is, once we have a little bit more time back on our clocks in all the different roles, we can now actually think of using a generative AI or just, let's say, the classroom to do other forms of education and things that we never have been able to do as much as we want to do. So if I talk to any instructional designer or learning designer, they're always unhappy about the fact that they never get to really roll out all the amazing methods of instructional design, or learning design, for all the activities they could do. Spend more times focusing on what different learning methods they could apply. Never spend enough time on the actual if you want learning structure, never spend enough time on case studies, or the assignments could be much better, much deeper, and so on and so forth. If I talk to faculty and subject matter experts, they always fall short on time in terms of sharing their research, sharing their practice, what they actually love, if they're teaching live, they really love the live interaction, right? But what they do is a lot of times it's just roll down a slide deck that they spend all that time preparing, then they're actually not really prepared to share anything else. So ultimately, the beauty there as a first step is that generative AI, by accelerating a lot of the groundwork, will be able to allow you, as an instructor, for example, to spend much more time engaging with the students. If you're able to build an entire learning experience relatively quickly. As a learning designer, then you can think about, what can we do beyond just the course production to make sure the learners have the opportunity to learn more effectively. And this is where I get particularly excited, because that's where it then comes the second part of AI. Now we can use AI to do something that is very much akin to the better way of learning, which is simulations, which is personalized content, which is activities that allow you to learn in a pseudo social environment, that is things that allow you to be much more applied, and so on and so forth. So we are. We're able to use generative AI to do groundwork, and then we're able to use generative AI with this neuroscientific background that I shared with you to create new formats of education.   Kaitlin LeMoine: It does occur to me, Frederick, as you're talking, and be curious as to your thoughts on this, that really you know you, you, we began this conversation with you, talking about the process of, for adult learners, of really assimilating existing information into or new information into an existing framework. And it occurs to me that, I mean, with, with AI, even if, like you said, right, it's maybe, maybe you've been aware of it for many, many years, and maybe now people have been aware of it for a year or so, adults are, I think, developing a new framework around AI. And so, how does you know? What is that work? What does it look like to apply your own kind of research, and the research around neuroscience and like related adult learning principles to this advent of this new age that we live in.   Frederick Wehrle: I think it's almost like a perfect storm on education and learning at the moment, and the pressure is as much on higher education, secondary education as it is on workforce development. I think the forces that are coming together is that leads to adult learners asking for more, asking for efficiency in learning, and are just not ready and willing to sit through content just cause, just because. Somebody put it into a framework. And it actually turns out, in my experience, that the educators themselves do not particularly enjoy producing the baseline content. They'd actually much rather spend the time and their time and energy on what is valuable. So in terms of AI and workforce development, I've two perspectives.  The first one is a lot of people need and want to learn more about AI and how that will impact their their work. And the second one is the environment. And AI enables them to learn anything that they need to move ahead and go further in their in their jobs, and stay ahead in their careers, or maybe even in certain situations, change their careers. So the first one how to learn about AI, I think the most critical part there is similar to what we're trying to do with that group of partner universities, institutions, organizations have to invest into, let's say, the experimenters in innovation labs and structures that will allow their organization to paint a picture of what it will look like in the future. Otherwise, you will always have that, that very strong resistance. And for you as a person, it is very difficult any individual kind of it's very difficult to to navigate the path of insecurity. And I think right now, with AI learning, there's so much out there. How do I stay informed? There's so much FOMO, and you feel like any hour you spend on that one thing, if you're interested in this, is a hour you lose, because then you didn't focus on the other thing. So you never feel on top of the situation. So I think there is a lot of work to be done by organizations to actually invest into people that are able to provide frameworks and an outlook of what it will look like if your organization is not doing that, and you are an individual, I always recommend actually going to events if you have the opportunity, be it virtual events or conferences or things like this in your field of expertise. For the last two years, all the places I've gone, anyone and everyone talks about AI anyways, it just gives you a much more concrete outlook on where things where things are going. And the third advice I give is actually take the time to watch at least the keynotes of all the large developer conferences that are going on throughout the year. So you want to hear the heads of Google and Meta and Microsoft and Apple, Nvidia and so on, talk about their releases and what's going to happen the second part, in terms of how you can use AI for your own education, I think there's, again, these two different aspects, from an organization or institutional perspective. You can use generative AI to actually create situations that make learning much more effective. So you can either use your student's existing mental framework and create content that makes them able to learn very quickly. Give you a quick example. Imagine you have 30 people that need to learn something, and instead of giving them one case study about a shoe factory, which nobody's working in a shoe factory, you actually allow the students or the learners to put in their job and cooperation into an AI or pre-configured GPT that builds you and each student an individualized case study, knowing that all case studies are based on the same learning objectives and the same kind of generic questions that they have to answer at the end, so every student will have an individualized, 10 page case study that really speaks to them, that they can understand very quickly, and then they can still discuss it with the instructor very effectively. The second thing is, you can use generative AI, particularly if you have the opportunity to do simulations, to create a situation where you can build mental frameworks. So if I want to train somebody up in something they have never experienced (nursing is an environment where this is particularly prevalent, or medical school as well), where you have to have these very, very long time practical and training periods, because certain cases just don't happen that often. I don't know a violent patient just doesn't happen every every week. So if you need to learn how that, how to handle that, you need to be able to experience it quite a bit of of times. And if you haven't experienced it, it's very difficult for you to learn the content of a textbook or so that explains to you how to handle this situation. So what you can do is you can create a simulation effectively. You simulate with generative AI, an interaction that is generated on the fly, where you just set some parameters effectively for the different actors in there, and you act out the situation. And you can interact with that verbally. And once you have that experience, then you can quote, unquote, hit the books and learn everything about it, and you will be able to learn it very quickly. Other example is project management scope out. This is an experiment I'm running at the moment, how to scope out a project. So what I'm trying to do is have a simulation with it's a startup that's called Convey. They have the possibility of you creating a virtual avatar, and that's based on a knowledge base, and you configure it however you want, and then the student just clicks on the link, takes on the role of a new member on the team, and that avatar is the boss of the department, you have to engage and talk with that boss and figure out what you need to scope out a new project. And the reality is, if you do this before learning anything about scoping a project, you will make every mistake in the world. You will feel terrible, especially if the character is made to be a little bit egregious and doesn't have much time, and it's like, okay, is this done? Do you need anything else? Then you really have an intense in five minutes, intense framework that is built through that experience. And then you read the text on how to scope a project, and everything will make sense instantaneously. And you can then even go back into the into the simulation. So that is actually interesting for for both the organizations, institutions, as well as the individual learners. AI has the possibility to personalize based on what you know and what exists in your brain, but also learning designers, particularly and instructional designers, can work with experts to create the situations you need to experience to learn more effectively.  So Frederick, while the world waits for you and your colleagues at the other universities to produce the road map from your perspective, what even more sort of practical steps can our audience take to become forces in keeping pace with the changing education and training needs of industry in particular, and and even be interesting to hear your take on like, I mean, part of what I spend a lot of time with is like playing around with different AI tools. Some seem to be better at this, and some seem to be better at that, and then a week later, this one's better at that, and this one's better at this. So can you offer just sort of very practical steps that people can take to be forces?  Kaitlin LeMoine: Recognizing they might change next week?   Julian Alssid: Yes. Thank you, Kaitlin. Frederick Wehrle: I think the biggest thing you can do, it's a change in mind set in, in how you approach your day and how you approach your work. Most of us are very much set in non-AI work processes. And I myself had to reconfigure my thinking, in my the way I approach any any problem or task, and think, how can I solve this with AI before I try anything else? And it's quite interesting to consider that. So to your point, there's different AI tools. I would absolutely recommend people, if they haven't done this before, they should definitely just take any LinkedIn learning or YouTube or Udemy intro to prompting, how do I use models, and get comfortable with using these models. But once you have that, I would not hesitate to use several models. I do as well, and try to have them at the very least, assist me, if not do the work for me. And the interesting perspective there is that I can configure the different bots so that they can take on different roles. So what I would recommend for anyone very practically is to have a virtual you so you effectively create a GPT or a Gem in Gemini that you configure to be equivalent to you, like your position, your context, upload examples of what you do, what you write, and so on and so forth. But especially give it the expertise, or configure it to be an expert in XYZ and have that role, because at that point you can actually use that as a second you to work with yourself, have it do things on your behalf. And the second thing that I use models very effectively with, and I know colleagues do as well, is when you use them as experts. So if you're in terms of designing, if you're doing learning design, instructional design, or if you're trying to develop your own skill set, you can give or create, either give the chat bot while you chat with it, or you create your own GPT or Gem that has a certain expertise, right? So you can create an expert instructional designer. You can create an expert subject matter expert in a field. You can create an expert colleague, and then you can actually engage, I mean, work with those experts to help you along the way with your tasks. In terms of keeping up into with the pace of the development. What that would allow you to do is to keep up as these tools evolve, as you mentioned, and that, for me, is almost more important than necessarily following the news or like being kind of on your social media, constantly trying or hunting after the latest information is to actually be practicing it, because then, through the practice, as the new updates roll out and so on, you feel comfortable integrating them into your into your toolkit. So these are, these are just the practical ways other things that that I do, and I think other colleagues do as well, in terms of keeping up to date, as I mentioned earlier, as conferences in your field, particularly, and configuring your, let's say, your news feed very consciously to follow things that that are relevant to your work environment, like a Google alerts you know where you can get the information that is very pertinent to your structure, but I felt mostly and most importantly, that's the change in mindset of using it first before you do anything else, being very comfortable that gets you very comfortable with it, that allows you to just be up to date and then seek out your community and the others that are excited about this within your industry, and try to connect on conferences, events, social networks, wherever you can.  Kaitlin LeMoine: It's really helpful to hear those very practical strategies, Frederick, and and it strikes me too, right, like the just how opposed, in some ways, the idea of, like a many year degree is relative to this type, this approach to learning these, these tools, which is really just like, kind of keep continuously up to speed and try things like, don't try to be the expert first, that just sticks out as pretty distinct with AI and kind of how to most efficiently learn these tools.  Frederick Wehrle: If you're trying to upskill yourself. And I think there's a lot of confusion out there right now. I think a lot of people have a lot of or procrastinate around using AI because it seems it seems complicated. It seems like a huge, huge task, and people read more about it than they use it, and especially like worried about your job, and how is it going to impact my job and so on and so forth. All of this gets quote, unquote. This is not really debunked, but all of this gets kind of resolved if you allow yourself to actually use it. I did it myself at a very early on, when ChatGPT came out, for example, and was readily available to spend quite a bit of time actually using it and realizing how much it can do what you're doing and how much it is better at your job than what you're doing in many respects. And so you panic. Then you don't sleep for two weeks until you have engaged with it enough, or two months. You reach a point of level of engagement with it where you see, oh, this is where my contribution lies. This is where my expertise comes in. This is what I will be doing in the future effectively, I might quickly say I always keep forgetting this, because it's never how I approached AI. Because of my background, there are a lot of people who use AI as a tool, so just know that that's not how it's built. If you're using AI as a tool or chat bot as a tool, you're effectively not using the user interface or the you're using it wrong in a certain way, because it was configured and built to be a conversational colleague and partner, right? An assistant. So it is made to understand and engage with you through full conversation. If you bark commands at it, it actually doesn't really know what to do with it, right? That's not how it was built. And so the output is not really good. And I've seen massive changes, and people like completely change their relation and engagement with AI, just from shifting their mindset to I'm not the expert, it is the expert, or at very least, it's a colleague, and I'm just treating it as if it was and I'm having a serious discussion with it. A colleague of mine always says he wants to challenge anyone on a panel, or anyone who is like giving a talk about AI or especially politicians and others, and ask them, Did you ever have a philosophical discussion with ChatGPT? And most of them haven't. And if you haven't, you probably have never actually discovered what this, what these things are capable of, and how you can engage with it. Kaitlin LeMoine: Well, with all of that being said and thinking to the future, how can our listeners learn more and continue to follow your work? Frederick, as it continues to unfold. Frederick Wehrle: I do usually post things that are interesting through my LinkedIn, so I'm very open. If people are interested, just reach out over LinkedIn. I do have a Medium account where I do put things from time to time, and I'm going to be publishing more, evidently, in terms of research papers as we as we move ahead with our framework, because we realize that that's where experimentation has to provide some output as well. We can't just keep it for ourselves. Julian Alssid: Well, thank you so much for taking the time today, Frederick, it's been really fascinating, and we certainly want to follow you and keep in touch. Frederick Wehrle: Appreciate that and very happy to be in touch as well. Thank you very much for having invited me.  Kaitlin LeMoine: Thank you so much, Frederick. Really appreciate your time today. That's all we have for you today. Thank you for listening to Work Forces. We hope that you take away nuggets that you can use in your own work. Thank you to our sponsor, Lumina Foundation. We're also grateful to our wonderful producer, Dustin Ramsdell, you can listen to future episodes at workforces dot info or on Apple, Amazon and Spotify. Please subscribe, like, and share the podcast with your colleagues and friends.