Future of Work: Will Degrees, Skills, and Careers Survive the AI Economy? | Michael B. Horn

Welcome to Starter TV. I'm journalist Kiran Shettri, and today we're talking to Michael Horn. He's a bestselling author of the book Job Moves and a thought leader in disruptive innovation in education. He's also an adjunct professor at the Harvard Graduate School of Education, and he co-hosts the top education podcasts, Future You and Class Disrupted. Michael Horn, thanks so much for being with us.

Thanks for having me.

To start, will you give us a quick overview of your career path and what you're focused on these days?

Yeah, so I, and I'm an accidental arrival, if you will, to the education field. I took Clay Christensen's class while at the Harvard Business School, and he's the father of disruptive innovation theory. And he invited me to collaborate on a book with him applying these ideas around innovation to the problems in public education. initially focused on K-12. That became my first book called Disrupting Class, and together he and I founded a nonprofit think tank called the Clayton Christensen Institute for Disruptive Innovation. Basically there, when I was leading it, we took his theories on innovation, applied them to public sector problems. I continued to focus on education. Then when you're in K-12, you realize, gee, to really solve these problems, you have to move into higher ed. Then you realize when you're in higher ed, gee, to really solve these problems, you gotta move into workforce. learning, because it's all a continuum and it's all dependent on each other. And so, over the course really of my 20-year career, you find yourself working in all of those spheres. And I've been writing books on the topics around the future. I podcast. I advise companies and universities and things of that nature. And really a smorgasbord, if you will, of a portfolio career today, where I try not to do any one thing, but I'll about a lot of these different things. And I do teach at Harvard now as well.

You've spent years exploring disruptive innovation in education and human development. What first pulled you into this mission of helping people discover their passions and reach their potential?

Yeah, I was so intrigued when Clay first started to teach me about these theories of innovation, because all of these things that seemed so counterintuitive that had happened around why organizations would succeed or fail and things of that nature just really didn't make sense. And so The opportunity to apply these ideas into the education field that is in sore need of innovation, that was what grabbed me at first. But I will tell you, the more I was into it, the more it became clear that my real mission is to help individuals build their passions, fulfill their potential, live lives of purpose. And education is a crucial lever, right, in that. And the way we have designed our education systems historically are ill-matched for the ways in which people in fact learn. And you would actually never design K-12, higher ed, workforce around the traditional ways of teaching and learning that we have if you were to try to build a system from the ground up to optimize individuals' learning. And so my question, just year over year of working in this space, was like, sort of become obsessed with how do you unlock this in individuals, and how is education a lever that helps every individual find their unique way to contribute in the world, as opposed to a system that's often weeding people out or trying to figure out who doesn't make the cut and is more obsessed with selectivity rather than actual development of individuals.

What do you think that traditional schools get most wrong about how people actually learn?

I think the biggest thing that schools get wrong is that their time is a constant and they treat every person's learning is highly variable, right? So you imagine yourself when you went into the 2nd grade, you know, your teacher delivered multiplication on the same day to every single kid regardless of where you were in your own mastery of the learning. And the challenge of that is clear, right? If you haven't mastered, say, double-digit addition, jumping into multiplication, it's not going to work very well. And that compounds upon itself, right? If you're doing algebra, we know that a huge problem that children face is that they actually have big holes earlier in their mathematics education or history. It's the same story, right? That these gaps perpetuate over time because we're so obsessed with moving at lock speed with the entire class regardless of the different background knowledge or mastery that individuals bring into any learning environment. So I'd say that big obsession around what I call the factory model of education one pace for all students, delivery at the same time for a mass of them. That's the biggest thing that I think traditional education systems perhaps get wrong, as opposed to what we've learned about the benefits of tutoring and mastery-based learning and things of that nature, where you're really optimizing for what works best for that individual at that given point in time, whether that's an individual experience or a group experience. That can be part of the equation, but it's not a set of questions we ask in traditional schools.

In your books, you predicted big technology-driven shifts in education. Now that AI is here, what developments excite you the most?

I think there's huge opportunities around AI on a few fronts. One, assessment, actually, right? To make it a much more seamless, within the education model itself, experience. I don't have to pull you out for your assessment and then get feedback later, which, as we know, is not a great way to actually improve learning. But it can actually be in the context of the actual learning itself. The AI can be alongside you understanding, is Michael showing fluency, mastery over what we're trying to teach him here? And really picking up not just on the knowledge, but also lots of other domains. You know, does he have the persistence it takes? How can we grow that in him? Does he have the executive function that we want to succeed in the workforce? How can we support him along that journey? So I think along a number of dimensions, assessment is a really exciting part about AI that is— we're just scratching the surface of right now. I'd say a second dimension that I think is pretty intriguing around AI, a lot of the early edtech products, while they were not bad at helping someone get in the reps, if you will, of learning, they were a little flat, a little uninspiring, a lot of multiple choice still, a lot of sort of narrow ways of engaging. a learner. With AI, you can imagine a much more robust set of conversations and ways to engage a learner in the learning experience. And frankly, with AI, it's a lot cheaper now to create really interesting learning experiences than it's ever been. And so some of the promise around simulations or virtual reality or things like that, that to be honest, I was skeptical of early on, I think make much more sense as AI improves and allows the creation of learning experiences to really ramp up.

Now that AI is here, what developments concern you the most?

Yeah, there's a whole list, obviously, of things that should concern us around AI and learning. The biggest one that, that is being talked about, and properly so, I think, is cognitive offloading to the AI, which is simply to say, I let the AI do the work for me. And as we know, the person who does the work does the learning. And if you're letting AI do all the work, you're not getting any of the learning, right? And so Now, same token, I do think that the expectation in the world of work becomes, how do I use AI alongside me to do even more valuable work and become more productive in my job? But what that I think means is that there's some things I need to learn without AI initially. Once I become more expert and fluid in those, then how do I use AI to complement me and create even better work product? But that's obviously a big concern that people have. The second one that I hear listed a lot is around cheating, of course, right? And that's sort of the corollary to what I just said. I'm using AI to do the work and I'm sort of cheating. I have a slightly different take on that, which is in some ways I think cheating is a byproduct of a system that is focused on sorting students out based on whether you show mastery or not, as opposed to a system that says, hey, we're here to support you in your learning journey and it doesn't matter whether you've gotten it yet. If you keep working at this, we're willing to keep working with you until you show mastery of this. And then the use of AI to, say, skirt learning, all of a sudden that becomes you're cheating yourself as opposed to trying to get ahead, right? And so I think changing the incentives in some ways, I think, sort of handles that. But I will readily acknowledge in the traditional education system, I think cheating is obviously the second big concern.

Do you believe that our current learning systems are equipped to keep up with the speed of AI change?

Yeah, it's a great question. I think so. What I will say is at the younger students, and we can just start there for a moment just to imagine the thought experiment, probably a lot of the stuff that you have to learn doesn't actually change that much in a world of AI, right? Like how to read, how to do basic numeracy, things of that nature. Some civics, question of history, things of that nature probably don't change that much. But as you get older, and I think middle, high school, college, working world in particular, the rate of change of tasks is unbelievable right now. And the way you use AI to do things is constantly changing. And so in my mind, the education systems that are built around a, hey, we'll create a curriculum, throw it over the wall 18 months later, you sort of go ahead on it. That's clearly outdated, and a traditional school system built around that sort of cadence is not going to keep up with the world of work. I think it's a huge call for much more experiential learning, much more opportunity to learn in context, real problems, real projects where AI is part of how I do the work, but it's really engrossing me in the environment themselves so I can keep up with what employers are demanding on the job.

You've said that education systems resist change because they're built for standardization. Are companies making that same mistake today, treating AI as just another sustaining innovation instead of a truly disruptive one?

Hugely so. I think actually, and it's interesting, you see it in the data as well, right? Which I think is that a lot of companies took licenses to AI outside of software engineering. Many of them have found that employees don't really use them all that much, or they don't change the work that much with them. It's sort of tinkering around the edges. And my sense of that is that companies have done what we always naturally do when we see a new technology come in, which is we use it to become slightly more efficient at the existing work, but not actually fundamentally rethink the tasks or processes themselves. And the real shift comes when you actually dig into the processes and say, what is the work? How do we do it? Let's reinvent it from scratch. That's where the real productivity and transformational gains come from, not layering AI on an outdated system, which is what I think we're seeing a lot of, not just schools, but businesses do as well right now. Very akin to how we treated electricity at the end of the 1800s, right? General use technology. The initial factories just said, okay, we'll replace steam with electricity, no big deal. Nothing really changed. And it wasn't until we actually redesigned factories from scratch and started decentralizing the electricity itself to machines and reinventing the work and the processes and the rate of work at different stations, that's when you saw the seismic improvements. I think it'll be the same thing with AI.

After the pandemic, a lot of schools and organizations rushed right back into business as usual. What opportunities do you think that they missed, and what needs to be reinvented in how we learn as well as work?

Yeah, the rubber band was real, right? I think a lot of us were hoping that the pandemic would force us to grapple with some pretty serious questions about how learning actually works. And to your point, people were so eager just to get back to sort of normal or status quo that they just rushed back into what they've always been doing. I think, frankly, we see the damage that that's done at the moment, because you see the disengagement in learning environments from kids, from college students, and so forth, is really down across the board. People are not showing up. They sort of have big doubts about the efficacy of the learning. A lot of the parents, in many cases, have questions about, really, is this what I'm sending my kid for? And so we're I think right now we're seeing a real disengagement and opt-out movement across the board, and that was the big opportunity that schools and learning institutions, in my mind, missed coming out of the pandemic, was how do we create a more engaging, a more efficacious learning environment? And rather than sort of go back to our one-size-fits-all models, how do we acknowledge People in different circumstances, they need different supports around them. They need different learning models. They need different modules, maybe in some cases, to really help them excel. And I think it's that one-size-fits-all thinking that's really been the mistake, if you will, coming out of the pandemic and the opportunity to really personalize for the different needs that learners have.

You often say that systems fail before individuals do. Why do you think that most corporate reskilling programs still look a lot like traditional classrooms when AI demands continuous modular learning?

Yeah, I think for the same reason we've said, it's a lot easier just to sort of keep your processes in place, not change the work itself. We used to say organizations, they develop processes over time, recurrent ways of solving problems. And what you notice in almost all organizations is they become very calcified. They're sort of like grooves in a factory floor that the machine can't get out of. And I think that's been true in learning and development processes in the corporate world as well, is we've treated them as like, well, this is the way it's been done. This is what we adhere to, compliance to these things, as opposed to stepping back and saying, how are we supporting the business? Right? Like, what are the real outcomes we want for the business? And if we start from the why and the outcomes, How would we actually rethink how we're upskilling individuals themselves around success toward those metrics? How would we measure that? How would we become a value add to the business as opposed to sort of a compliance effort in service of how we've always done things? And I think that's been the big miss again that corporations have had, which is corporate L&D has basically said, okay, AI is something we'll just fold into what we've always done. As opposed to really rethinking the work. And how do you become a value add and almost something that the CFO says like, oh yeah, Chief Learning Officer, I can see how you're contributing to the bottom line. You're not just a cost center because you actually demonstrably improved productivity by doing something really different.

When AI initiatives flop inside of organizations, how much of that failure do you attribute to weak learning infrastructure and adoption issues versus the technology itself?

The model almost always is more important than the technology itself. And so if you have a flawed model, it doesn't matter how great the technology is, it's not going to get the results you want. And so almost always, in my experience, the learning models and the learning processes in corporate L&D environments, that's the thing to blame, not the technology. And I just, I'll give you a quick example of this, right? You can imagine you have this amazing technology that personalizes learning. It assesses your mastery of a concept and serves you up the right lesson at the right time to help you keep making progress. Maybe it's in the flow of your work as opposed to something you step out to do for a couple hours a week or something like that. And say you introduce that. In an environment where the learning and development still happens on a particular cadence. We have your 2 hours, you know, every month where you have to pull out, go into this room with 20 other people. We're all going to do the same thing. It doesn't matter your level of mastery or your level of understanding of the skill or whatever it is. That personalization then is completely lost, right, in that implementation. It can even get worse than that because I might conclude Wow, this thing is trying to coach me. But then when I go into my workshop or my in-person session or whatever it is, it's systematically telling me it doesn't care what I've actually mastered or not because we're all going to do this activity instead, or we're all working on this compliance-driven thing instead, or whatever it is. And so then I might actually even disengage from the technology because I'm like, well, it doesn't actually really matter. They're not really looking at those results. They're really more about me checking the box on this thing rather than learning from it.

Many companies invest heavily in learning programs, but they see little real behavior change. What are they fundamentally misunderstanding about how adults learn in fast-changing environments?

I think the big thing that they're misunderstanding often is that processes have to change alongside the learning behaviors, right? So it's not enough just to upskill the individual. The system itself around them has to then encourage those behaviors and not be disincentives toward them. And Clay Christensen used to always like to say, there are not dumb people in the world. There are people working in dumb systems that cause them to act in dumb ways. And I think that's the really big takeaway is that a lot of this rubber banding to the way it's always been or sort of reversion to the mean, that's because the processes and system encourage that. Like the ways you interact with other— with your work colleagues or other departments at work and things like that. are not encouraging these new practices or these new skills themselves. But if you haven't attended to the process change in the system, you're not going to see these huge benefits nor individuals really using these new learning behaviors. It's almost a truism, right, that like it's just sort of what we incentivize is going to get done. The other thing I will say is to really change behavior, You have to really practice that thing in context over and over again. What we know from a lot of learning is that just because I can show it or do it within a very contained environment does not mean it's going to transfer to another environment unless there's really scaffolding and practice around that to make it very intentional. So I can see the commonalities and like, oh, When this problem comes up, this is how I'm supposed to think about and the things I'm supposed to do. Okay, now when I go into a less structured environment, are those same scaffolds and supports around that to really make this work?

Are most corporate reskilling efforts today genuinely preparing people for AI-driven roles, or are they mostly about signaling progress without any real transformation? Yeah,

I think most corporate L&D efforts today are still about just like checking the boxes of, have you done the AI module? Have you thought about ethics? Have you thought about compliance? Things of that nature. And to be clear, some of those are not unimportant, but I don't think they're the larger context. What's the work you're doing? Are we putting you with a project team? Maybe you're all trying to solve this problem in a different way and leverage AI. really, you know, in-context learning that really thinks about the process. Now, to be fair, I think there's sort of 2 parts of this, right? One is just building in employees the basic knowledge of what can I do with AI that I couldn't do before? What are the guardrails and so forth? And then like step 2 is, okay, now that I have this knowledge, how would I actually rethink the work itself? And almost giving that as the mandate So you can have this grassroots innovation as opposed to a sort of a top-down view of skill development.

In Choosing College, you questioned the ROI of traditional degrees. Do you see industry certifications, apprenticeships, and skills-based credentials starting to replace degrees as the main signal of capability in an AI economy?

I'm more measured on this one. What I would say is that there are a basket of college degrees that still have incredible value if you complete. And that's like the really big thing, which is roughly 40% of people who start a 4-year college degree do not complete college. If you're one of those, the benefits of going to college are probably negative because you're probably taking on debt and getting no benefit because you don't get the degree, which is the big signal in the workplace. If that describes where you think you might land, you shouldn't be going to college, right? Any one of those other paths you just named— apprenticeship, perhaps getting real work experience, jumping in and having a series of internships and so forth to get— or trade school and the like probably makes more sense. There's another set of degrees, 23% of them we know have a negative ROI even if you do complete them. Buyer beware around those, I would say. But if you're in the 60% group that isn't taking those low-value credentials, you're probably in good shape going to a college, right? And so I think the bigger I think your answer is it's nuanced. That if you're going to complete, you're going to get a degree from a place that gives you a lot of social capital and connects you with a lot of people in the region and alumni and things of that nature, probably going to be valuable. If you're going to a place that's not really going to create those connections, or you're not going to complete, or you're going to take on debt and it's not going to justify what you get out the other side, then you should be really looking for these alternatives around certifications, apprenticeships, and things of that nature. But even those have buyer beware cautions around them. And I think the bigger message to the individuals is do your homework. Don't assume just because it's a this, it's going to be great for you. Make sure this is really in line where you're set up to succeed, right? Like, I want to do the work, I have the conditions where I can do the work, and I can see other graduates from those programs really excel once they get into the workplace.

So you talked about this 23% of college degrees that are not really worth it. Can you tell us what those degrees are?

Yeah, a lot of times these are around high pay or, you know, high tuition. but low-salary things. So it might be cosmetology, for example, or certain social work degrees and things of that nature are really where you see the negative ROI. States have made it very difficult in those environments, I will say, because there's licensure requirements. So if you're going to go that path, great, but really try to have very little to no expenditure on your end if you're going to go that path, because you don't want to be out a lot of money for something that doesn't pay something to justify the cost.

And what about the college degrees that you said if you do complete the 4 years and get that degree, they are worth the investment by and large?

Yeah, those tend to be in areas, uh, in, in-demand fields— STEM, business, uh, healthcare for sure, uh, and also, uh, from more selective institutions, right? Where, uh, if you get into one of those institutions where people are clamoring to get into, those tend to have good value, uh, or at least not be negative. And they tend to put you with a group of people who's really going to help you get ahead. And one of the other things we know from research is people get jobs not just based on a baseline of knowledge, but really on who they know. Social capital is an incredibly important ingredient. According to some estimates, up to 80% of jobs are filled by someone in your network, which says networking and building your social connections is actually an incredibly important part of your development and your skill development that we sort of don't tell young people about. But it's the number one way you get a job. And I actually think what we're seeing in an era of AI is the importance of network is going up, not down, because what we know is happening is that people are using AI to apply to jobs online for them on their behalf. If you're on the corporate side, you have an applicant tracking system that is using AI to filter out resumes. And so effectively, you have AI applying to AI. No one trusts either side. And the better way to go is to actually just say, hey, let me call you up on the phone. This is the role I'm looking for. Do you have a couple young people or a couple people in their careers that maybe I should talk to that would be good fits for this? And you're like, yeah, I've got a couple people that would be great. Let me put you in touch with them. That's how jobs tend to actually get filled, and I think that's more the case in the future, not less.

You've talked about careers becoming more portfolio-driven in an AI-powered world. What does a strong career portfolio actually look like, and how do people build one proactively?

Yeah, I think a big part of careers of the future is that we're not going to just do one thing. Right? And the careers are clearly going to switch more frequently than they did historically. We also see more people like me who are in the gig economy doing a range of things at any given point in time. And what I would tell someone is when you take that first job, say it is a full-time job, have a side hustle as well where you're developing some other skills, you're able to showcase another part of your personality because the The truth about any job you have is there's no perfect job. There's no dream job along every single dimension. Every job has its element of suck. So use that to sort of complement yourself, allow you to scratch other itches that you're curious about, but also round yourself out better with a range of skills that make you more valuable to employers in the future, make you more valuable to do different kinds of work as things get valued or devalued in ways we can't always predict in the future. I think that's the trickiest part of right now, right, is we actually don't know, you know, no one who's being honest can assuredly tell you this is the thing that's going to really matter in the future, right? And so you having a range of things that you're developing and a range of skills and work contexts and so forth that you're doing work in, and you can show results from that in your portfolio, and people that are alongside you that can vouch for you,

You've worked with places like Minerva University and Guild Education, and you've studied innovations in countries like Vietnam and Korea. What are some of the most forward-thinking learning models that you've seen that traditional institutions haven't yet embraced?

I'll talk about the Minerva University one because it's my favorite. Students literally live on 4 different continents for their college experience, and the learning occurs in 2 places. Number one, it occurs online, but in synchronous seminar-based formats where they're literally extremely active on the learning. And that's one of the biggest ingredients that we know about learning is, again, the person that does the work, they do the learning. And so the more active I can get you, the better. And these seminar-based experiences have you constantly debating, asking questions, being put on the spot to defend answers and the like. And it occurs online with an incredible amount of analytics being used for the instructor. to really saying like, gee, Michael's slacking at the moment, or that didn't really have deep evidence behind it. Let's drill in a little bit more with him to really push his thinking and see if he can demonstrate mastery of these things. So that's one part of the learning. And then the second part is extremely experiential in the different countries in which people are living, where you're doing real-world projects in context with employers, civic organizations, and the like in these countries. You're developing cultural fluency, you're developing a network around the world, and you're developing real skills in context of meaningful projects where you're applying what you've learned online in the environments themselves, which I think is really valuable. Again, great example of active learning, great example of meaningful projects, great example of connecting it in context to the careers and sorts of things that people will be doing in the future.

What are the biggest structural barriers preventing large organizations from truly redesigning their learning and talent models in an AI-first world?

I think some of the biggest is that it's really hard to design the plane as you're flying it, if you will, right? So we often hear that as a cliché, but if you step back and think about it, it's true because essentially you have things that you need to do every single day. And you know you want to redesign how you do learning, how you do training, and so forth. But, oh my gosh, this employee calls up and they have a problem with X, or, you know, such and such learning is occurring now and you got to be there to support it, or there's a meeting with the CLO and you've got to be there on task. And so what ends up happening is that the urgent and immediate priorities always crowd out the important, but not immediate, not urgent sort of fires, right, that you have to fight. And what you learn from our research on innovation is that if you really don't have someone whose full-time job it is, sit there thinking about what are we designing for the future, that's my important and urgent, and I have no responsibility to the existing processes, it's just really hard to prioritize these big fixes. And so you really want someone to own these new things that you're trying to build, these new capabilities, so that you're not wedded to the old. And look, you know, the reason the fly the plane as you build it thing always gets laughs is you would never imagine a pilot actually is the one building the plane. You have a separate group of engineers who are doing that, right? Same thing in healthcare. Like, the doctors are not coming up with the life-saving drugs or treatments. There's another group of people that are doing that. And so I think that's the big takeaway is that in education and training and learning, for some reason, we expect the instructors almost to be doing both, which doesn't make a heck of a lot of sense when you break it down or think about other fields.

Through your podcast Future You and Class Disruptive, you've spoken to so many educators, founders, and leaders. What are some conversations that fundamentally changed how you think about the future of education or work?

So many conversations have shaped my thinking. That's why I actually love podcasting because it constantly challenges you and changes your mind as you learn from folks around the world doing interesting things. I'll name just 2 so that I can answer your question if that's okay. On my Class Disrupted podcast, we interviewed John Danner who had founded a charter management organization called Rocketship Education. And now in the world of AI, has founded something called Flourish Microschools. And there he's using AI to be like a constant listening companion, to be assessing people on all the, what he calls super skills around executive function, agency, persistence, all those things we've known are important, but it's been really hard to measure in any meaningful way. And the AI is giving you feedback. Yeah. and so forth on your performance so you can really improve over time. That, I thought, was fascinating use of AI that you're not really hearing in the conversation that much. That started to open up my head for what we could do with this technology. The second one I'll name is a more recent one I had with my co-host on Future U, Jeff Selingo. We had someone on the podcast, Michael Beddersworth at Texas State Technical College. And their funding model is such that 100% of the dollars that they get is based on the economic income that their student— former students are generating for the state of Texas. So if a student is not gainfully employed, then Texas State Technical College gets no money from the state for that individual several years later. That sort of 100% accountability model Jeff and I were talking about, and Jeff made the point to me, it's not that everyone should be moving to that sort of a model. It's more that we should have clarity around what are the outcomes we're incentivizing for any given learning model, how would we measure them and pay for them, and then create the right incentives and metrics around that so that, you know, your community college is not trying to look like Harvard and pretend that it's, you know, pretend that it's selective and elite like that, but instead can be recognized for excellence around its mission. And it gets paid more if it actually delivers on that mission. And same thing with, say, your 4-year liberal arts. They're not trying to look like Texas State Technical College, but they too could be paid in line with their mission and value. And make more money if they're excellent on it. So we sort of stop comparing things that are fundamentally apples and oranges, but we also allow for excellence and prestige to accrue in all these different categories that I think we need in the future.

If you're looking 5 to 10 years ahead, do you think that traditional institutions, universities, and corporate L&D will evolve fast enough, or will entirely new learning ecosystems replace

My instinct is we'll be looking at entirely new learning ecosystems over the longer run. I think that's certainly true over 10 years. 5 years, I'm a little foggier because I think things are actually moving a little bit slower than we think in corporations with AI because of the dynamics we've talked about, that people are treating it as a technology problem rather than one of managing the work and the processes itself. And so the adoption is somewhat slower than we would have expected. But I think what we're going to see in 10 years' time is a whole series of new business models, new processes, new organizations coming in that are doing the work, and then a whole set of new learning providers, or maybe we'd say complementary learning providers, right, to those models that are really custom designed for this world in which we're living, in this world of rapid skill change and the like. And it's a very different ecosystem as a result that will really marry each other, if you will, in that future.

If we're having this conversation again in 10 years, what would need to have happened for you to feel genuinely optimistic that education and workforce development have become more equitable and personalized in an AI era?

Oh boy, that's a great question. I think the thing that I would want to see is genuine focus on the outcomes and a clear sense of, hey, when we're using this learning module or when someone enters this college degree program, what is success? What are the outcomes we want? And how do we know? How do we measure it? And let's look at that, right, appropriate to the program. And can we almost see, has there been meaningful movement on those outcomes in ways we measure consistent with the vision, mission statements, and so forth of those different providers at each of those different stages. That's the big thing I would want to see. And I think AI should make it easier to measure those hard-to-measure things that we know are important but have too often sort of fallen out, if you will, of the equation, but we know have really mattered. And so that's the sort of thing I think I would be looking at.

What is one thing that you'd want someone watching this interview on Sardar TV to walk away thinking about? Something that is probably what they would not get from a typical tech blog on AI readiness.

Wow. Uh, I think the biggest thing I would say is the model matters more than the technology. And so, you know, we've said it before, but I think people just don't believe it, that if you really want different results, you got to redesign the model itself. That means the processes, the how you do it, the what are the goals, and then the technology is something that supports that. And the really cutting-edge tech is going to be custom-made for these new models that really question first principles. If you haven't questioned the first principles, I don't think you should expect to see big changes in terms of learning.

Well, I want to thank you so much for being with us. A fascinating conversation. Thanks for being with us, Michael Horn. We really appreciate it.

Thank you.