Podcast: Processing AI in Education Out Loud

Get stories like this delivered straight to your inbox. Sign up for The 74 Newsletter

Class Disrupted is an education podcast featuring author Michael Horn and Futre’s Diane Tavenner in conversation with educators, school leaders, students and other members of school communities as they investigate the challenges facing the education system in the aftermath of the pandemic — and where we should go from here. Find every episode by bookmarking our Class Disrupted page or subscribing on Apple Podcasts, Google Play or Stitcher.

In this episode of their miniseries on artificial intelligence in education, Diane Tavenner and Michael Horn reflect on what they’ve learned. They discuss how AI offers unprecedented access to expertise, but also highlight concerns about its effectiveness for young learners. Throughout, Diane and Michael grapple with skepticism, optimism and the practical challenges of embedding AI in educational systems, while looking ahead to what meaningful, student-centered innovation could look like.

Listen to the episode below. A full transcript follows.

Diane Tavenner: Hey, Michael.

Michael Horn: Hey, Diane. It’s good to see you.

Diane Tavenner: It’s good to see you, too. I’m excited to see you in person in the coming week. But, you know, and, and maybe we’ll just jump right in because I think people know we’ve been doing this miniseries about AI and I’m going to jump in because I’m very excited for this conversation today. We have been talking to all these folks in this little miniseries, and you’ve been doing a better job than I have of sort of just listening to them and letting them talk and not, you know, sort of interjecting your opinions and feelings. You know, that’s a little bit hard for me. But today is our day where we get to do that. And so for our listeners, you know, sometimes you, you write to us or you tell us that you like kind of overhearing us talk to each other. And so this is that episode.

We have not talked to each other about all that we learned and discovered from these conversations. And so we’re calling this kind of our out loud processing episode where we’re going to go through and just process through. What do we hear? What do people say? What are, what were we thinking? What’s our takeaways? And then we’ll come back one more time and organize that and put that into kind of a final season episode and a final miniseries episode where we’ll pull out the big headlines and the takeaways. But today it’s going to be pretty raw conversation.

Michael Horn: Today is going to be raw, up close up, personal, all of the different demons in our heads. And if we miss something, send us a line, tell us what we missed. Because we have looked back on some of these conversations. Some of these conversations, I suspect, Diane, it’s going to be more like, wow, I, this, this one thing has been really burning with me and I had to address that, but I sort of forgot some of the other points. So don’t be shy about pointing that out to us. We, we really have enjoyed, I think it’s fair to say, these conversations because it’s really opened up to a lot of different perspectives. And I think there have been elements of truth or insight in every single conversation. Even where we disagreed on certain elements or whatever else to me, it’s just sort of like revealing the whole elephant, if you will.

Diane Tavenner: I couldn’t agree more. That’s probably a really important place to start, is with some gratitude. Thank you so much for the people who came to talk with us and share with us. It’s been incredibly valuable for us. What we seem to be hearing is it’s valuable for other folks. And so let’s dive in. I’m, I’m going to take sort of that … I know for a fact that both of us, this has stuck with us.

It was from our first conversation with John Bailey and I think it was, for me, it was such an eye opening conversation. The idea that John is so clear that what AI provides is an expert on every subject in every person’s pocket. And this idea that we now have this expertise just literally at our fingertips. And not only did he say that kind of at a high level, but I think what was amazing about what John did was he literally gave us these concrete examples of how he’s using AI as an expert in his life. And they were like so many different examples. And then I of course went on to like to look at more of them and read more of them. And you know, he’s probably the most creative person I’ve talked to about how he’s really using AI as an expert. And so that one’s just sitting with me.

I don’t. And I’ve heard that from other people too. How did that one strike you, Michael?

Michael Horn: Yeah, so similar thing, Diane, which is, I think you take away, OK. Google gave us access to the world’s knowledge. This gives us access to expertise. I think that’s like a really interesting distinction. It lands for me. I think frankly, the thing I took away, or one of the things I took away from the episode we had with Siya from, from OpenAI was less the views on education and more how she actually uses the tool herself as this personal assistant to guide her learning agenda, to help her figure out what to learn and on and on. Made me feel somewhat inadequate as like a human in terms of all the things that I could be doing with this. I think I’ve forced myself to increase my usage in certain ways since then.

Diane Tavenner: Have you watched yourself changing at all because of these conversations?

Michael Horn: Yes, yes. So I will say and, and I, I’m curious if you’ve had the same thing, but number one, I search on Google a lot less than I used to.

Diane Tavenner: Almost never for me,

Michael Horn: So the only reason I don’t is because I have access to the Gemini Advanced, I can’t remember what they call it, but the advanced AI search feature, right. Which has a lot of the same qualities as ChatGPT or I guess a lot of folks use Perplexity for search because of the way it answers. Right. Your queries. But yeah, in general, I am not using semantic search really at all. I’m almost exclusively using AI to try to understand certain things. I will tell you, I was trying to get a much deeper understanding of the nursing and healthcare shortages across the country recently. I had Google Gemini and I’m blanking on the exact product, but it’s their research product.

Create a, like a, it’s like a five to seven page basically briefing for me on it. Super interesting how it did it and it, and it resolved one of the challenges I have, which is when you sort of just do a raw search, you get like, oh, by 2036, this is the projection. And I’m like, no, I want to know now. I want to know by specialty and where I like. Right. And you can get that now. And, and it’s, it’s really interesting Diane, what about you?

Diane Tavenner: So similarly, the only time I find myself going to a traditional search is out of habit. And then I get there and I’m like, wait, why are you doing this? You’re going to get much better information. I’m using the paid version of chat right now.

Michael Horn: Yeah, that’s what I largely use. I should.

Exploring ChatGPT Usage Trends

Diane Tavenner: Yeah. Well, it’s interesting. One tip I’ve gotten from, you know, sort of insider is to, to actually cycle through them and use the different ones from time to time and see what you think. So I’m going to try to push myself to do that and not just fall into a single habit. Although we’re not alone. You know, in the last couple of days the numbers have come out about the number of people across the planet using ChatGPT and it’s extraordinary, like unprecedented. We might get to that in a little bit. But yeah, I find myself pushed by John in a lot of ways to just push my thinking on, wait, do you really need to be doing this? I keep asking myself that all day now, wait, do you really need to do that? Can, can an expert do that for you? Or, or in some cases things that I thought I couldn’t know.

I’m now saying, wait, I think that might be possible. Like how could, how could we get to that? And so I, I feel like it really, these conversations have pushed my behavior and change and with positive results.

Michael Horn: I mean, yeah, it’s super interesting. I’m curious, Diane, if you’ve had this question come up which is so we’re learning how to use it and I feel like I’m still very much learning right. How to use it in this way that increases productivity, efficiency, and the realm of what’s possible for us or me to accomplish. You as well, it sounds like. I guess I’m curious as, like, you think about that in the educational context.

Diane Tavenner: Right.

Michael Horn: If you’ve had reflections about, OK, so what does this mean at different levels of education? Where have you gone with that?

Diane Tavenner: Well, I think that’s where it starts to move out of my own personal excitement and curiosity given where I am in my life and whatnot, into the reality check on K-12 education. Because very few young people who are in high school or middle school are experts at anything. You’re just not an expert yet. And my, when I was listening to John talk and I was listening to Siya talk, I was like, but you guys are experts. So you have this set of skills and knowledge that enable you to use this tool as an expert for you. But novices oftentimes don’t know how to take advantage of expertise, so it’s not accessible to them or open to them. And what young people are doing in their lives and their learning is fundamentally different from what we are doing in our lives and our learning.

And so one of my big questions coming out of those conversations was like, OK, great, but what happens to, you know, the folks that I’m really focused on, the young people and, you know, ages, adolescence and in early adulthood, how did they, given what they’re doing on a daily basis and thinking about and trying to do, how does this concept of an expert work for them? And I would argue it doesn’t in the sort of raw form that we’re accessing it.

Michael Horn: So. Yeah, yeah, no, I, how do we always end up in the same place? OK, so, so there’s this notion, right, in learning sciences of novice versus expert learner by domain. And then there’s the second sort of, you can create the two by two. Right? So you have novice and expert on one domain or one dimension, and on the other you have unknowing versus knowing. So, right. You’re an unknowing novice. You literally do not know what you do not know. And you just have like very, very little expertise in this.

Diane Tavenner: Yeah.

Michael Horn: Then you become a knowing novice, meaning you actually start to understand the realm of things people in the field and domain do and all the things you don’t know.

Diane Tavenner: Right.

Michael Horn: And then you become a knowing expert. Right. Is sort of the continuum and you still have a sense of like, I know how I learned to be an expert and I know the sets of things that I did and you tend to be a really good teacher when you’re a knowing expert. And then you become an unknowing expert, you start to automate 75%, I think Bror Saxberg tells us, of, like, the things that you, you, you, you sort of do on a daily basis, the underlying skills and things of that nature, and you just sort of forget about it. Right.

It just fades into the background. It’s automated. Right, right. And so what’s interesting, I think, as I reflected on this, is from my perspective, um, the, and I like the way you just said it, the raw form of these tools is probably most useful in the knowing, like, circle of that. So I’m a knowing novice, but I at least know the questions to ask. Like, I have a certain set of foundational skills and knowledge in the domain that allow me to, like, use the AI as that personal assistant to help guide my learning and like, you know, be curious to interrogate an answer it gives, et cetera, et cetera. I think it’s also true that probably when you’re a knowing expert, that it’s really useful for boosting your job performance. And, my hypothesis, Diane, is that this is why we’ve seen so many studies come out that suggests AI is most helpful to the lowest performers in the world of work and least helpful for the biggest experts.

Right. And those, I think are your unknowing experts, is sort of my guess. And again, and then on the unknowing novice side, I think it’s probably not super useful either, or frankly, sometimes maybe even misleading. Right. In certain cases. And so I think you need really highly curated learning tools. Right. If you’re going to be using it for individuals like that.

Now this gets. Maybe I should pause there for a moment because we could talk about equals. Yeah, yeah.

Diane Tavenner: But I love it. I love thinking about it using that framework. And I, you know, I am very concerned about the unknowing novices because by definition, that’s who we’re getting and serving. I mean, that’s, that’s a natural state for them to be in, in their life and their developmental journey. And so, and what I, I think you, you know, I am not a fan of chatbots. And from the very beginning, you know, when people were getting so excited and their first kind of conceptualization of how this could be used is we’ll just basically take the little box and we’ll just put that little box all over the place, everywhere. And then people will just come and they’ll just know Ask the box.

Questions. And then it’s solved. Everyone’s just going to learn. And in, in my mind, that’s a, you know, that’s a chatbot, and that is not going to work for the unknowing novice. They don’t even know what to say or what to ask. And this is proving to be true. I, I’m, you know, in a lot of conversations, looking at a lot of data where people have essentially chatbot data for young people, and you will not be surprised to learn that they write weird things.

Improving AI for Learning

Diane Tavenner: They write in, you know, like, short, incomprehensible things. They’re not asking questions. They get frustrated in there. They’re yelling at it sometimes because they feel like it’s supposed to help them, but it’s not helping them. And so what I have a little hope about is we’re a little bit further along now and people, I think, are starting to be able to imagine beyond a chatbot. So how do we actually, I think this is where you’re going, like, how do we actually use AI in products for younger people, unknowing novices, and even the emerging, you know, folks to help structure their learning and help to teach them, but not just to put this open box there for them that they have no access to. And so there’s a little bit of promise on the horizon as we get a little bit further into it and people start to process and think about how it can be used. But, but to me, that’s the, that that is one of the big risk and I think one of the reasons that you see the folks who are very skeptical about it, and we had a number of them, we talked with a number of people,

Michael Horn: A lot of skeptics on our show. Yep.

Diane Tavenner: Yeah. And so. Yeah, yeah.

Michael Horn: No, I think that’s. Yeah, that all lands for me, Diane, where you’re going. And, and I guess from my perspective, it does point to something which I think was true in the era of Google as well. Which is it. It’s not the case that we don’t need to learn knowledge. Right. Or at least what I would call foundational knowledge. And I thought Rebecca Winthrop was really good on that concern.

She sort of said, I’ve been the skills person and now I’m worried we’re going to forget about the knowledge. And it goes to something we talk about all the time, which is like, we have to get away from the tyranny of the OR in this education world. This has to be an And conversational. And I think Foundational knowledge is really critical, right? To being able to use these tools in, in ways I think people are really interested in creativity right now turns out to be creative, you actually need to know something and then to be able to break the rules, right? And like interdisciplinary is really important then, but. But you do need to have some foundational knowledge.

Diane Tavenner: I wanted to go here next, like the direction you’re leading us because I think both Rebecca and Jane surfaced a really important conversation about skills and knowledge that you’re bringing up. And I would argue, you know, folks who’ve listened for a long time know that I’ve always organized skills, knowledge, and then habits of success and in the habits realm is curiosity. And so I’ll talk a little bit more about that. But what I’m interested in. One of the things that I notice that often happens in these conversations around learning is that skills and knowledge get really. They’re not distinguished from each other. They are put into the same category or bucket. I think it might be worth just unpacking a little the difference between skills and knowledge and habits and for, for a conversation for education.

Because like you just said, Michael, knowledge is, let’s say for the purposes of our dialogue and our conversations, it’s, it’s the stuff. It’s like the names and the definitions and the dates and the, you know, the, the theories and, and those sorts of things. And then concepts are sort of a little bit of a bigger idea of knowledge. Skills are the things that literally you practice and can improve upon and that go, you know, are more universal and stretch across and use the knowledge, if you will. And so just to be very concrete about that, a skill being, for example, to a high level skill is to effectively communicate or to analyze or to solve a problem. These, they’re people’s favorite skill that they like to talk about is critical thinking. Critical thinking actually has a whole bunch of skills.

Michael Horn: Many, many skills. Yeah, yeah, right.

Skills, Knowledge, Habits: Learning Framework

Diane Tavenner: And many of those that I just named you, those are the big high, you know, domains, and they have multiple dimensions. But think about things that you can actually practice and improve. And so if we call back to Jane’s conversation in the writing center and her as a teacher of writing, I mean, skills, skills, skills. So much of what she was talking about was skill development, right? Knowledge. I mean, people have been worried about knowledge ever since, for forever. Because, you know, can you just look up a fact or can you just look up a date or something like that? And, and, and then I think the third category, and then I’m curious about your thoughts about this, that I like to distinguish is this idea of what we would call habits of success. And this is sort of a big catch all for everything from like, how do I emotionally regulate myself? I’m calling back to the good work that the building blocks framework that sort of identifies at least those habit, what I call habits that are related to school success and learning success. So everything from, you know, can I emotionally regulate myself? Can I have, you know, can I be in relationship with others? And then all the way at the top of those building blocks has always been my civic identity, self direction, which, you know, has been a huge center point of how I think we need to structure learning.

And curiosity. And curiosity has always been fascinating because super hard to measure. No one really knows how to teach it or if you can teach it. But what I think is happening right now is illuminating the critical importance of curiosity and how our system of learning and education has sort of rung curiosity out of young people. And it might be the most valuable skill habit in this.

Michael Horn: You’ve anticipated me again, I think when a student asked me, you know, one of the students at Harvard asked me recently what I thought was the most enduring skill. But, but habit is how, you know, you and I have generally classified it like it would be in a world of AI. And curiosity was the answer that I had for a couple reason. One, I think when you are getting answers or interacting with whatever the form factor is, being able to interrogate it and knowing how to ask and not settling right is going to just be like baseline importance. Right. And then two, though, I think like in a world in which the rate of change is accelerating in terms of the world of work, this curiosity as a gateway into learning and upskilling, et cetera, et cetera, et cetera becomes really, really important. So on multiple levels, I think curiosity is critical. The other habit I’ll name Diane, from the building blocks goes down to if like I want to say it’s not the bottom layer, but I think it’s the second layer, you’re going to correct me, but which is self awareness, I think is the one or something like that, or self advocacy or something like that that you all have and you can redefine it for me if I mess it up.

But I think this is like really knowing yourself and like the strengths that you bring, frankly, not just your strengths, but also like what you suck at, the things you don’t want to do. And I don’t think I’m going to be stronger about that. Schools today do a very poor job of helping individuals learn around their self awareness. And like what, what, like what? You know, what superpowers do I bring in? What are my weaknesses? Where should I walk away from things? I get why that happens. We don’t want to give up on an individual too early from developing something that could be a strength.

Diane Tavenner: Yeah.

Michael Horn: And I think as you get out in the real world, you realize that life is lived with your competitive advantage and the things that make you unique and, and not trying to remediate your weaknesses constantly. And so I think in an era of AI where, look, AI is going to be the new expectation in the workforce. Right. Like you, you don’t use it. What? Is going to be sort of the question. But you can use it to really effectively craft your career in a way that you couldn’t before, because now you can let it do the stuff you don’t want to do. Lean into the place where you can add unique value. Well, that requires self awareness.

So those are the two habits, Diane, that I think are very, I mean, I think obviously all the habits have enduring value, but the curiosity and self awareness, I think are really important.

Diane Tavenner: I totally agree with you, Michael. And I think there’s a couple of other things to, like, illuminate here around why I think we don’t do a good job of sort of nurturing young people into being, you know, really aware of themselves. Well, I, I just don’t even think we try to do that.

Michael Horn: Yeah, I don’t think that’s been a goal. Right. Of the schools.

Diane Tavenner: Right.

Michael Horn: And just so people do not misunderstand us, like the report card you got a C in social studies, like, that’s not what we’re talking about. Right.

Diane Tavenner: So, no. And, there’s a couple of things going on there. One, for all the stuff we’ve talked about over the years on this podcast, we actually don’t give young people and their families very honest information. And by honest, I mean information that they can truly understand and interpret that tells them and gives them feedback about where their strengths are, where their weaknesses are. The grading system that we have is woefully inadequate in terms of giving actual feedback. And our testing system is, quite frankly, as well. You know, when, when, when I get a report of my child’s state testing and I have a hard time reading it and understanding what it says, you know, that though, like, this is not working for families, you know, we’re not telling them what their young people are good at or not good at. And to be fair, one of the challenges with that is there’s a base level of skill and knowledge that I think all young people need that it doesn’t really matter if you’re good at reading or not.

We need to get you to be good at reading. Like you need to be able to read. And so there’s not kind of a picking and choosing. There is as there will be later. But I want to jump in on this idea of. Because this is a lot of the work that we’re doing right now about like knowing yourself. And I think the approach that we’re taking is just from like working with David Jager and then learning scientists like we will be good at work, our work, our career, our vocation. It’s pretty simple.

Passion Fuels Career Success

Diane Tavenner: If you like it, you’re probably going to be good at it. And the reason is because if you like it, you are more curious, you are more willing, you are more interested, you want to do it more, you practice it more, you get better at it. And it’s a self fulfilling prophecy. And so one of the activities that we ask young people to do is to really look at the things you will be doing in a job or a career every day. What are the top 10 tasks that you’re going to do day in and day out and are a big part of it. And then very honestly self assess do I like doing those things or not? And it’s really shocking how it’s hard to figure out what people do in a job every single day. It doesn’t really come through in job descriptions or in most of the tools that young people are given to think about career and jobs. And it’s actually a thing they really wonder about, which is why they want to talk to people who are in the job to ask them what it’s really like.

So they have an intuition around this. But that like assessment of and that, that just realism about do I like doing this? Because if I don’t I’m not going to be very good at it. And so I should pursue the things I like doing. And, and I think that gets translated into people saying follow your passions, which is a wholly unuseful thing to say.

Michael Horn: Unuseful. Yeah, completely. Yeah.

Diane Tavenner: So, let’s make it more concrete for them. And so I, I would argue, you know, that’s, I’m with you on that. I would say the skill that goes into building is reflection. And we don’t as a general rule spend nearly enough time teaching young people how to truly reflect and then use that reflection to propel them forward.

Michael Horn: Yeah, it makes a lot of sense. I want to stay with this, just because the point that Jane made specifically right around this was that the process is what’s important in writing. It’s not the product. In that case, it sounds like for those of us in New England who have, you know, who had Bill Belichick as coach here for however many years, it’s the process, not the right. And we became a big mantra here. I think that’s true probably in the Bay Area with the Golden State Warriors, too, but. But like focusing on the process as the learning. I think this is interesting also because reflection is built into that.

And I’ve heard some. I want to try this out on you. I’ve heard some people say, you know, so. So I think part of Jane’s answer was like, I still need you to do the writing. And maybe it’s. Some people I’ve said in class said that AI is not there and see the process. Others I’ve heard say, like, do the writing. I believe you’re going to be using AI to do it, but I want to see the questions and prompts and things like you’re asking it to do as a reflection on the process and, like, how it changed, you know, how it changed the final product, if you will.

I’m, like, curious as a, you know, someone who taught writing, like, what you think of that as a mechanism and does that make sense to you?

Diane Tavenner: Yeah, it does. And. And you’ve taken me down another path I want to ask you about, because there’s all these legislator legislation, state bodies now that are trying to pass AI legislation, and it’s this full range. So I’m curious too. To go there with this. Texas is top of mind for me right now. So as a. As a former writing teacher and, you know, I am who I am, so you’re not going to change that.

Like, I think it would be silly to try to banish AI from the writing process. And let’s be clear, that’s what some people want. So there’s a.

Michael Horn: Let’s be. Let’s be clear. I teach at Harvard, where there’s a policy that unless your instructor explicitly says you can use AI, the default is no. I think that’s insane because these people are going to go into the world with an expectation to do it, so we might as well make it intentional. So I’m on record there.

Diane Tavenner: The further education, taking one more step away from the actual real world and the world of work and saying, you know, What? We’re not going to prepare you for that. We’re not.

Michael Horn: Yep.

Diane Tavenner: And so. But we’re, we’re aligned there. So how would I, as a writing teacher, think about it? Well, I mean, in my experience right now, I’ve watched young people who are not skilled writers try to use AI to write something for them. And first of all, you can tell immediately, number one, that they didn’t write it. And two, it’s not good. It’s not very good. And so I think that’s where I would start is just being really talk about feedback, being open and honest and like, let’s actually dissect what happens when you just try to put in a basic, simple prompt and get something out that is like. But quite frankly, this is also what school does, is just put something on paper and turn it in versus actually building a skill.

AI Bans

Diane Tavenner: And so I think there’s a big opportunity now for, for great teachers, great instructors. And I actually think we heard Jane talk about some of her strategies here to help young people understand a tool that is now available to them and will be in our world, and how they can use it to not only build their skill, but improve their products and their outcomes. But that is going to require a whole new set of skills from them and muscles that they are not using and flexing in school right now because they’re incentivized to not do those things there. And so I think it’s very exciting and hopeful and optimistic, and that’s why I get very disturbed when I. I mean, there’s literally a bill in Texas right now that could very well pass that is going to say something to the effect of, you know, teachers in the state are forbidden from using AI in any teaching and learning.

Michael Horn: Yeah. I mean, I’ll be. Yeah.

Diane Tavenner: What. What is that?

Michael Horn: Right, Right. Every product now has AI, so. Yeah.

Diane Tavenner: Yeah. Yeah. Well, and then I’m like, you. I think who it was that some university was saying. Yeah, they’re. They’re literally going back to blue books.

Michael Horn: Yeah.

Diane Tavenner: And exams. I’m like, really?

Michael Horn: Yeah. I mean, I hear the same thing. I, It’s. I didn’t know about the Texas bill. I, I will be very consistent on this one, which is, I do not think making policy at the level of inputs ever makes sense. And I feel that way. We had a whole set of shows about the science of reading, and we were super clear about, you know, the importance of using actual, you know, like, actually following the research on this. Right.

And so forth. And I don’t believe in policy at, you know, banning certain curricular materials because I think it stifles innovation when you see leaps forward. And look, if we want to pass, you know, measures to create professional development so that the people coming out of schools of education actually know how to use these tools, teach science, reading, use AI, whatever, I can have that conversation. And I just, I think it’s a blunt axe. Even when I’m in favor of the spirit behind it, shall we say here, I’m not in favor of the spirit behind it. I think it’s a blunt axe the wrong way. It’s the same reason I, you know, feel that way around mobile phones as well.

I want schools to have the ability to take them away and not have them when it does not suit them right on the ground. I don’t want a policy criminalizing the teacher that found a good use for it and one person in the school disagreed and then all of a sudden it’s, it’s a thing. I, I just think that’s misguided.

Diane Tavenner: Well, that’s another bill in Texas too, so we’ll see what happens. I want to stay a little bit on this thread, but I want to go to something that, you know, you and I are both, I mean, our work has been steeped in personalized learning. And so you know, Ben Daley or Ben Riley.

Michael Horn: Yep.

Diane Tavenner: Ben Riley joined us and another one.

Michael Horn: Another of our friends, Ben Daley we’d say.

Diane Tavenner: But Ben Riley joined us and you know, pushed pretty hard on. He believes that the promise of personalized learning has, is sort of overdone adjudicated, it’s failed. And he believes that, you know, this, you know, the hype about, well, I shouldn’t put words in his mouth. Everyone got to hear him. Let’s say he’s a skeptic. He’s a self-described skeptic. And he did bring up this idea of personalized learning. It also came up, I think in our conversation with Julia.

Michael Horn: She was more optimistic about it, but yes. Yeah.

Diane Tavenner: So a number of folks talked about the idea of personalized learning and it seemed to me that there was kind of these two different like either like see, personalized learning is like, you know, this is just gonna, AI is going to go the way of personalized learning. It doesn’t really work. It doesn’t really personalize or oh, AI is actually going to. We’re still on the journey towards the vision of personalized learning and AI actually helps accelerate us in that direction and improve the possibilities. I’m kind of sticking up, you know.

Michael Horn: Sure. The extremes yeah, yeah, but, well, but no, I think there’s something to the way you did it though. Right? Because what I hear a lot of advocates saying is like, well, now we finally have the technology to do all the things we had imagined 10 years earlier. As though the technology is going to sort of automatically understand, you know, like what you’ve mastered and your working memory capacity that day based on what you ate and so forth. Like, and somehow deliver the perfect lesson at the perfect time. Which I think is essentially right. Sort of that techno driven vision.

Rethinking Personalized Learning Paths

Diane Tavenner: Of personalized learning, which was never sort of my vision and what we do, but that there is that version of that. And so it’s got, it just got me thinking about like, oh, OK, where are we with personalized learning? And what do I think about that? And is this, are we on the same journey or pathway or have we hit a sort of fork in the road? And does this change my perspective? I think maybe if you’re, you are one of those techno vision people, it probably does. It feels like a huge accelerant. For me I think it’s a powerful tool to continue down the path of realizing the vision that I have for young people, which is that people have always confused it as being like an individual kid on a computer, and that’s never what it was. It’s much more how do we use technology as a tool to prepare young people for the real world, for real life, for real skills. And it’s a very powerful tool, if used well, to do that. And then also what I get excited about now is how it can actually structure our system of education and create efficiencies and opportunities that I think have never before been possible. So I’m very optimistic about what it can do, probably more on that latter part than on the first part .

Michael Horn: Say one more beat. Like when you say in terms of what it can do on the system part, what does that look like in your mind? Or you know, sort of simple sketch? What does that look like?

Diane Tavenner: Well, like I’ll give you an example, you know, that I’ve been pushing myself to try to. OK, if I could design a school from scratch right now, what would it do? And that’s because I’m a nerd. That’s fun for me. So that’s like a pastime. And one of the things I imagine, let’s just talk about how a family might engage with school. So, I’m going to give you this utopian vision. But like, what if, you know, periodically you sat down with, with Your family and your girls. And you were able to say, you know what, over the next couple of years.

And you, you did this like, with technological prompts, like, over the next couple of years, what’s most important to my family about what my girls learn? And they’re, they’re different from each other. So I suspect you would have different things where you and Tracy would be like, well, this is really, these are my top goals over here. And these are my top goals over here. And of course we would scaffold that for you and we give you a menu to choose from or a list or, or perhaps some, you know, but we would ask you as a family, like, what’s really important to you? OK, like, check all the ones that you care about. Check the ones you don’t care about. And then this is like my analogy of how is school like, ordering a sandwich, you know, like, and then we would go through a series of prompts to be like, OK, well, let’s get into your family. Like, what does your schedule look like? Like, do you, you know, do you want a day, a week with your girls at home with you and your family? And they go to the building four days a week? Do you want to come at like 10 because of the way your family schedule is and go later? And you know, and I can imagine people starting to have a heart attack right now as I’m talking, like, oh my gosh. But I think if we really, truly went and could ask and understand the circumstances of every family, literally AI can do what humans can’t do, which is it can go and crunch all of that, and you can ask it to help you design what would be possible within the parameters of what the school can actually offer.

And not every family has to be on. Everyone arrives at 8:30 in the morning and everyone leaves at 3:00 in the afternoon. And one day a week we leave at 1:00. So, you know, like, we don’t have to do that anymore. We have technological capabilities that could actually bring a whole community together and meet their needs in a personalized way.

Michael Horn: I think that’s really interesting. So many thoughts going through my head as you say this. One, I think the importance of context of the individual. Two, look, not everyone will get every, like, we might be out of romaine lettuce that day and there’s trade offs, right? But the point is, and this is what’s always driven me nuts about the world of personalized learning is the word personalized learning as a noun, and implying that like there’s one way to like, oh, I’m personalized and you’re not.

Diane Tavenner: Right.

Michael Horn: Whereas instead of seeing it as like a verb or. Beth Rabbitt, I thought, did a good job in this chapter she wrote for us in this new edited volume, School Rethink 2.0 of like. It’s a series of strategies you can do to better meet learners with what they need next, right in their, in. In their learning journey. And at that level, like, I, I just, you know, Ben Reilly is a big, you know, he learned a lot from Dan Willingham, the great cognitive scientist. You know, Willingham talks a lot about, right. Like, if you put something in front of someone that is way outside of their, you know, zone of expertise, proximal development, if you want to go right, that they will get frustrated, tune out, if it’s too easy. And I see this, like, I see technology tools right now.

I will not name companies, but they’ve sort of bought into the, oh, it should be all whole class. And I see that, like, yes, they’re following the learning sciences, say, around reading and the importance of knowledge to build understanding, to do the skills right, et cetera, et cetera. But because every kid is doing, like, reading the exact same book from a teacher who’s following a script, right? Like my cousin, excuse me, my kid’s cousin, she’s like, I read this three years ago. This is the most boring thing. Like, I literally want to jump out of the window. And she disengages, right? And I suspect the truth is on the, on the other side, that, that to me is insane. And so less. It’s like magical, technocratic, personalized learning and more, hey, this is a strategy with a set of tools.

We have to better come closer to meeting different family needs. When I hear the structural one you just laid out, my mind goes to the, you know, the world of education choice, right, where we’re starting to see that with education savings accounts, where these are the experiences that families are constructing. I think what’s difficult right now is, like, we know how hard it is to arrange summer camp as a parent. We did a whole episode on that. We’re kind of asking parents to now do that the entire year. Yeah. So to your point, how does AI maybe services maybe different kinds of bundles, right? Like, so you, you walk into Subway and we’ll go with your analogy, right? And like, they kind of tell you, hey, Here are the 10 best combinations of the stuff. But, like, if you want to custom build it, you can.

Yes. I kind of think that’s like, we have a Rebundling along these different, like the most common, if you will, set, set of customizations or personalizations.

Diane Tavenner: I just want to pull a couple of those threads and just be pretty explicit about them and why I think this is important and addresses some of the big challenges we’re seeing right now. So one, I think so many of the battles we see across the country right now are people who, and I’m talking among parents and you know, we’ve talked about school boards and all those things are about people who want a certain thing for their child. And because the school only does one note for everyone, if the school’s doing the thing that they don’t want for their child, they then therefore try to change it for all kids.

Michael Horn: Yeah.

Diane Tavenner: And this is causing massive, you know, fights and battles. It’s very cultural. I’m going to keep picking on Texas today because I’ve just spent a bunch of time digging in on them. I mean, they are taking back control over the curriculum so that at the state level they can really control. And this is very much about cultural, like what young people learn or don’t learn. In response, I think a lot of this. And so to the extent that we could personalize at least parts of education, I think it tones down some of this. Like what, what is true for my child doesn’t have to be true for your child.

And they can both get what they need without compromising the other child. So, there’s a benefit there to it. And then I would say, I think you’re absolutely right. There’s a ton of people who are really worried and against ESAs and vouchers and things like that because they feel like it’s the unraveling of our civic society and we won’t have people together., you know, building society together will be, you know, further in our, our camps or our bubbles and whatnot. And I think that, you know, the vision I just painted for you of how folks might get into school, I agree with you. There would be trade offs just because you marked it on your sandwich sheet. You know, that day we happen to be out of pickles.

Like, it’s just not going to work. There’s no pickles. You don’t get those. Sorry. You know, but I think people could handle that and accept that more in the good of the, for the good of the community and the group if they felt like they had some control. And I think the problem with our choice system right now in America is it’s so blunt. It’s like, you can pick a school. That’s it.

And that’s such a massive, we need a scalpel, not a big blunt instrument, you know, like.

Michael Horn: Yeah, no, I agree. So, I think models like this are emerging. Right? Like Alpha schools. It’s a private school originated in Texas, so this is a bright spot. And they had the two hour learning model, which is essentially as I see it, Diane, like what homeschoolers have done for years, which is like we learn the academic, you know, basically content and some of the skills right. In two hours. And then we like to go out in the world and do real world immersive experiences. They just are using the AI in a very, I think, developed way.

Diane Tavenner: Right.

Engaging AI-Powered Learning Tools

Michael Horn: To offer that two hour learning sequence. And then frankly, this is the other piece of it that I think is going to be important if we’re going to need to think about motivation a lot more. So if we build these curated AI tools that can work with the unknowing novices, we’re going to have to connect it in ways that get them engaged into actually wanting to learn these foundational knowledge and skills, which we should be doing anyway. But Right. Like I think and, and, and we’re not like that’s the evidence of the chronic absenteeism, disengagement, et cetera, et cetera. But I guess I think that’s like, we really need to think about how to create meaningful engagement. And I think this notion of, hey, you can learn sort of your nuts and bolts, your foundational stuff that’s critical much more efficiently and then get to do much more engaging work because there’s a connection between them.

Diane Tavenner: Yeah, yeah.

Michael Horn: Should be, should be part of that answer.

Diane Tavenner: 100% and I think purpose. And so that’s why you go back to personalizing people’s purpose. Like why are you here? What I mean it’s, it’s to your work, Michael. Like what are you hiring school to do for your family? Right, yeah.

Michael Horn: By the way, that is the best question. Yeah, sorry. When people ask me what should I do about my kids school, Tracy tends to jump in the conversation because she said he’s going to get towed in, in the weeds. Let me just tell you, like, what is the thing that your family can’t or, or, or isn’t able to do that school can do for you? Right.

Diane Tavenner: Like what are you, what job are you hiring it to do for you? And, and it will be a different answer for different families. So I want to keep us going.

Michael Horn: Sorry, we’ve deviated perhaps.

Diane Tavenner: I do want to acknowledge that I’m thinking, thinking about this, this infrastructure benefit and this is what Julia was trying to get to, I think, in her points. And this is a vision that she sees. And so it’s interesting to go back and think about some of the comments that she made about it. Michael, One of the things that surprised me honestly was that basically everyone we talked to, like these AI isn’t for kids who are under 18 right now.

Michael Horn: Oh yeah, that was fascinating. Were you surprised? Were you surprised by that?

Diane Tavenner: I was surprised by it. And so, you know, at least now the adults who are thinking about this, working on this, we’re very much focused on the adults that are teaching or doing things for young people, but not kind of a direct use for young people when we push them. They did talk about, you know, while it could be embedded in products or maybe, maybe not.

Michael Horn: Well, I mean, it is, let’s be honest, right, everyone.

Diane Tavenner: Yeah, but that was shocking to me and I don’t know why, why did that shock me?

Michael Horn: Why did I, I was super shocked as well. I mean, I think obviously, right, Privacy and some of the really detrimental impacts of social media and these consumer companies are clearly part of what’s going on here. I think that caution is good. I do believe, despite what I just said, that I don’t believe in bans on, at the policy level of mobile phones in school. I do believe a lot of the John Height research, I find it compelling that social media, specifically on the smartphones, has led to a bunch of antisocial and problematic mental health outcomes and disengagement. So I think that’s a lot of what’s going on here, Diane, is sort of my guess. And like they also, I think we need to also be honest that kids are using these tools.

Like we are not a huge screen time as, you know, household. And my kids have certainly had experience with ChatGPT. They have certainly used it for many things. That is certainly how they search at this point when they want to prove a point to me about something.

Diane Tavenner: Well, and you know, my, my kiddos are, you know, a decade older than yours and they’re early in their career and it’s, it’s. Well, one of them, it’s what he does all day, every day for his career. But the other one is literally working around the clock to make sure that he is becoming expert at using it as an early career professional because he feels like if he doesn’t, he’s going to be, you know, pushed out of the job.

Michael Horn: It echoes, you know, what Matt Siegelman from Burning Glass Institute has found, which is that AI is actually used more in sort of marketing, communications, professions like that than actually even sort of coding heavy parts of the workplace. Which is interesting. It’s not what, it’s not what I would have expected.

Diane Tavenner: Yeah. Yeah. That is fascinating. You brought up two things that I’d love to touch on. So. And we can decide where to go. First one is this idea of like AI being embedded in products. And I actually think it’s worth us sort of surfacing.

What does that even mean? And what does that look like beyond a chat bot, if you will? What are we seeing? You know, you know, still feels like it’s still early, but things are moving so fast it’s not early. So anyway, that one. And then the second one is this idea of, you know, Julia brought a very real fear about the loss of, potential loss of social connection. And so I want to come back to both of those. Where do you want to go first?

Michael Horn: Oh, we can do, we can do embedded products first. Embedded product for 200, Diane. So, yeah, what do you, what do you, what are you seeing?

Diane Tavenner: What are you embedded?

Michael Horn: Yes. We have not yet, we, we are not yet. We have not yet been replaced by AI doing our voices. But what are you, what, what do you, what are you seeing out there in the market? As frankly someone who’s building and I think using AI yourself in the product, but not, but not leading with that.

Diane Tavenner: No. And so maybe that’s the good place to start. I see a couple of different categories. So one is there’s folks who literally jumped out of the gate immediately and labeled their company, you know, AI. AI is in the name of the company somewhere. They are AI forward, they are AI first. They are like and what I find with those is many of them weren’t even sure what product they were building, but they knew they wanted to build an AI product.

So it’s sort of like a. AI in search of a product kind of origin. And yeah, I think, I think what I see over there is like people who kind of started with a chat bot in some sort of realm and then they’re maybe like evolving it over time because I think they’re probably getting feedback that great. A chat bot in a specific area is not that super helpful. But let’s, let’s name some things like that. There’s like companies that are like, we’re going to provide, you know, AI driven mental health supports. So we’re going to train A model to essentially be a counselor, if you will, that you know, can engage with and interface with young people. There’s AI tutors obviously in reading and math, you know, all across the board.

So, I see that as one category. I think the second category is, I hope it’s a category, I think it’s where I sit, which is having a very clear vision of what we want to do and why we want to do it with our product. And then we, on a sort of case by case decision grid, decide if AI can be useful or helpful for this particular part of that. If so, how are the trade offs worth it? And then decide where we’re going to strategically use it in the product itself and then also in our, in our work. And I would say that the in our work part is much easier and kind of a no brainer because there, there’s an efficiency tool and things like that. So, so that’s, I do think there’s a category of that. And then I think there’s a lot of people who are existing products and existing companies, you know, this is the majority, they’re not startups and they’re having to figure out how they get an AI strategy with the products that they have built that didn’t necessarily have any element of that.

So I don’t know, do those.

Michael Horn: That feels like a pretty good way to categorize the market to me as well. It’s interesting in our opening episode we had this dichotomy of student facing versus teacher facing. And as I hear your reflections on that, like that sort of cuts across those categories in interesting ways. I think both are like interesting ways to view the market at the moment for different reasons. And, and, but, but the way you just categorized it I think is largely what I’m seeing. I would say the market in terms of funding startups is moving away from the first category being the thing. You know, there are a couple home runs in that space, right? Magic school that is used by millions and millions of teachers, right to lesson plan and dramatically make their lives more efficient and by the way, for them to personalize for kids that maybe they were struggling to reach. So, you know, really cool boomed out of the box.

I think you’re right. The majority, I think, are now increasingly sitting where you are, which is how is AI an enabler of something that we’re trying to effectuate here, right? And then I think what you see is that, yeah, the large incumbents, if you will, they are using AI in different parts of the product stack to enable different things in different ways. Right. And in line with the way that they currently come to market or operate. I don’t think that they’ve used it to overthrow right what they’ve done. It’s more as a amplifier of what they’re doing.

Diane Tavenner: Yeah, so I lied. Let’s not go to the social connection yet. Let’s stick with that right there for a moment. Because one of the big things I keep wanting to ask you about is we’re having these conversations like, OK, step back to your work around disruptive innovation or innovation. And we’ve had these conversations before of where an innovation sits. Like walk me through where you place AI in.

Michael Horn: Yeah, that’s great. OK, so I, I think I’ve said this before on the podcast, but like fundamentally, AI is a technology enabler that can be used to sustain, which is what we just outlined the existing companies have been doing or to disrupt by fundamentally creating something that is dramatically lower cost, more accessible. Right. And serves people who don’t have access, which is what you’re trying to build. Right. In terms of this guidance and sort of understanding who you are and charting your future. Right. System or tool.

And so that. So again, it’s sort of. Yes.

Diane Tavenner: So AI is, the big category can be right.

Disruptive Educational Innovation Emerging

Michael Horn: Can be both. Right and, and so but here’s like an interesting thing in that which is back to the conversation we had earlier of the education savings accounts world and not just school choice, but education choice and like in many, you know, in 63 different flavors of ice cream or whatever it is. If like that is growing share, I don’t know how big it is, but that’s going to be a very different distribution channel into market with the eyes potentially helping you right. Figure out like customize for you. Are the existing companies, like, they don’t that those aren’t their customers today. This could be, I, I guess, Diane, where I’m starting to think is like, if I, if we truly move into that world, right, I as a family can stay in the district school, but like I might be then like losing out on anywhere from 7 to $16,000 in an education savings account. And now all of a sudden it has a cost to me to maybe take this. And so now like we can actually move into a world where there’s actual disruptive innovation of schooling, not just disrupting class.

Michael Horn: Right. For the first time in our country’s history, since 1930 or 40 or something like that. And then like that opens up all sorts of disruption opportunities, that’s into the market more broadly. Right. Like right.

Diane Tavenner: Technology I hadn’t thought about but this idea that you think families don’t put a price tag on like a public education? They do about it and so now when they’re staring at well like I get nothing over here if you will, because it’s not quantified in a dollar figure. But over here I get to spend some amount of money I had not thought about.

Michael Horn: I don’t know, I’m super curious is what I will say. Diane. Right. But like it it if you stop holding public schools hold harmless as most of the ESA, maybe all the ESA still do, at some point that’s not going to continue. Right. Like at some point you’re going to have to do what they did in charters and take money. At that point if like families are going to have real trade offs that they’re wrestling with, I think and making choices for their kids. And if there’s a series of services or products or things like that.

Right. That like dramatically help you get what you need for your kiddo in the context of your family environment, that opens up like a mind boggling number of possible disruptions in the market, I guess is sort of the bigger point. And AI look, it is not marginal, zero marginal cost, like sort of how we thought of the Internet before, which itself wasn’t because of distribution. But like it it is you are able to build stuff with dramatically fewer resources than you were. And so if you’re starting from that point and you’re not contending with an incumbent that has a huge advantage in terms of distribution in this world, what does that open up? I, I, I think it could open up a lot of things and, and incumbent, both district and incumbent, like large curriculum players. Right. So yeah.

Diane Tavenner: Right. What’s coming to my mind right now as we started this podcast, as people have heard us say a thousand times at the beginning of the pandemic, because you and I thought that the, it.

Michael Horn: Could be this opening yeah.

Diane Tavenner: Could be finally the thing that really broke it open and disrupted education as we know it. We both admit we were wrong about that. So here we are, season six, still hoping, but now talking to you about this and this is why I wanted to ask you that question is AI, I mean you seem to be making a case that it could.

Michael Horn: Well I think it’s part of the narrative. Right. And so it’s like, I actually think in an interesting way though, the pandemic will be part of the narrative too. Because it dramatically increased the number of families consider these options. And I think led to, yes, ESAs, etc were bubbling, but it dramatically increased the openness. Right. Or the desire of families for that adoption. And so I think all these things come together and I, like, let’s, I’m not ready to make a prediction, but I think it opens us up to something that could be very different.

Um, yeah, like a very different moment. Put it, put it that way.

Diane Tavenner: I think what’s interesting about that, when I think about the scope of history and, you know, my kid is a big history buff, and so he’s, he always says, like, what gets lost when people look back in history is that they think something happened really fast. But if you really look at the history, it happened over 60, 70 years. And those were kind of painful years for the people who were living through them. Right. There’s a lot of, like, churning and disruption and whatnot. But then we look back and we’re like, oh, that happened in like a minute. You know, and so I feel like living through, you know.

Social Connectivity and Dislocation

Michael Horn: so the dislocation is, it’s part of it. It’s uncomfortable. Maybe that’s the gateway into the Julia question of, like, how will it impact social connectivity? I’ll just jump in with my thoughts on that for, for what it’s worth, Diane, which is I, so I believe her fear is real. I’ve seen, but I’m, you know, I’ve seen some people say, like, really is in response to the episode. I’m, I’m actually not concerned about it emerging, though, in an education use case, as in, I believe the reason sort of the individualized, personalized learning version of the world didn’t come to pass and would never come to pass is like, people, like, being with other people and sort of that experience is really important. And a tool, for example, that is giving you career guidance to stay in your lane is going to be really useless if it doesn’t connect you to real individuals at some point in the journey. And the reason for that is the way we get jobs is through our network.

Diane Tavenner: Right.

Michael Horn: By conservative estimates, over 50% of jobs are through your network. As high as 85%. Right. No one really knows, but it’s somewhere in that range. So a tool that does not at some point push you out into the real world and connect you to real people in my mind, is not going to work. And so I, I hear Julia’s fear of, like, well, we may have the wrong metrics and policy around these things. Yeah. But at some point like people are going to be like, this thing is useless, it is not connecting me to real people.

And so I’m less worried in the education use case. But I think she’s right. In the commercial use case, these companion bots in effect, right. Anthropomorphic, as she says, identities of AI, you know, are, are, are, are a real concern. And so I think she’s right to worry about it. It’s the part of the social media narrative into this one that I think we should be worried about. I don’t know where it goes. I, I will say I’m, I’m not against those, you know, real world simulations and things of that nature as part of the learning ecosystem.

I do think it does ultimately need to connect into the real world of real people as part of that continuum. Right. And so AI, I think can be a really useful tool for creating the individual simulation where you learn to work something in the privacy of your own home. And you, yes, like you are less afraid to ask a question because of social, you know, in my case, like what an I banker do when I was a junior in college. Right. Like I would have done that, use that. And at some point then it has to connect you into the real world in a real world experience. So, like I’m less worried about her thing in the educational context, but in the world of loneliness and social media and AI filling that void, I think that is a very serious concern and it will ripple into our world of education and impact our schools.

Diane Tavenner: Yeah, that all resonates with me and where I go with it because, you know, I can’t help it as the practitioner is, well, what does that mean for our work? And for me it reinforces the idea that, and what I think the promise of personalized learning is, which is we actually give more time. In a well designed, like elegant design of a personalized learning experience, there is more quality time for people, adults and young people, young people and young people to be engaged in meaningful, authentic work. You know, what I’m going to call know myself work. Like the work, there’s nothing more important than knowing who you are. Building a healthy, developing a healthy identity, developing a healthy self. And like this is what we could be doing in education through like, go back to what David Jager talks about, like what do young people care about? They care about status and respect and there’s very precise definitions around that, but in their community and in their peer group, and it comes through earned respect. Like I do something that contributes to this group. I make a meaningful, you know, contribution that’s respected by others and therefore I am, I’m given sort of status in the group.

And that all happens when you’re doing project based learning, real world learning, you know, coaching, reflection, self development, that’s the stuff we should be doing together in person. And then personalizing the knowledge acquisition and some of the skill development so that I can come and access that and be a part of that group. I think in an elegant personalized learning model. And to me that is prophylactic against the fears of what would happen in the commercial world. And quite frankly, the fears that exist right now around social media and the damage it’s done if young people were building healthier identities outside of that world, that’s, that’s how they can resist, you know, the, the perils of social media.

Michael Horn: It’s well said. I think nothing is inevitable in this part of the landscape. And this is why I think it is so important that the educators, education entrepreneurs in the world that I just sketched out of a world of ESAs are super intentional about creating those opportunities. Those opportunities could be in the school communities where kids are coming together. It could be in connection with the community organizations around you. And I think there’s a, you know, there’s this big debate going on of like, hey, we need more career technical education schools. They’re really expensive to build. And then someone says, oh, but they’re cheaper than sending someone to college.

That’s a misfit for them. And you’re like, actually there’s like a kind of interesting middle ground of like leverage all the infrastructure around us of employers and companies and community organizations, et cetera, where people can actually plug in. And you’re right, like that foundational work that maybe will be a little more solitary around foundational knowledge skills so that you can actually come in there, you know, being able to contribute in some way. But those are all connected and I think we have to be super, super intentional about it to ward off, sort of ward off the dark side of that story.

Diane Tavenner: Yeah, we scheduled a long time because we knew we were gonna go long.

Michael Horn: Can I make one more point, can I make one quick other point on this? Yeah, just it’s one of the things that Ben Riley hit over and over again was that AI does not in fact think like humans and therefore will be less useful than we think it is because it does not think like us. To me, that’s a bit of like a, there’s a word for it that I’m not. It’s not coming to. Truism is not the right word. But it’s sort of like, yes. It does not think like us.

That doesn’t mean it can be. Cannot be useful to us. Right. And so that’s the parsing I would love to pull is like, I think it actually can be very useful as long as we understand the intentionality behind it and we’re clear around that. Not in a pie in the sky way or not in sort of a technocratic, oh, we just mix in technology with existing systems and models and poof, it magically works. I don’t think that will happen. Right. I do think we have to have intentionality with what we’re doing, with what the outcome we want from it.

Does it map on to learning sciences? Does it map on to how we build creativity? Curiosity, or at least not stamp curiosity out and sort of the schooling forms, if you will, that exist in the future. So that’s just like one other thing that I thought was worth reflecting on.

Diane Tavenner: It is worth it. And I might just say, and hold me to it. This will be my last thing I’ll say. But you reminded me that one of the things that struck me from these conversations, and I think it’s because we’re still really early, but like, everyone is looking at AI through their particular expert lens and we didn’t get a lot of, like, broad conversation outside of people’s expert lenses, my hypothesis is because it’s still really early and people are just trying to make sense of it. And of course, you first make sense of it through how you see the world and what your work is. And certainly that’s what we saw with Ben, you know, and his kind of views and. And you know, what felt pretty narrow actually, you know, but then through all of our guests, I think we just saw kind of how it is relevant, specific to them. It’s made me try to push myself and think, oh, am I being really narrow? And how can I think more broadly and to be on the lookout for people who are thinking about it outside of their own specific domain.

But maybe this is where we need to sit for a while.

Michael Horn: I think to your point, like, there’s so much moving every single day, you know, like there was that study out of Harvard on the physics class, right? They had done the flipped physics class, however many years ago. It produced better learning, continued to do so, as I understand they used a tutor for active learning. It. It produced better results than people said. Well, it could be the Hawthorne Effect, right? It could be. It’s narrow foundational knowledge. Does it really do this? I don’t know. Like, it’s promising, and we have a data point on it, and it was a real RCT.

Let’s. Let’s watch. Right. Does it solve engagement? No. It doesn’t solve all these other questions?

Diane Tavenner: No.

Michael Horn: OK, so let’s just say what it does, and let’s keep thinking about it. No silver bullets. And it made me so appreciative of the series we’ve done here because I didn’t know what we would learn from our guests. I feel like I took something away from every single one of them that altered how I think about the landscape here in meaningful ways.

Diane Tavenner: I completely agree. And we sort of bring our processing session to a close. I will say I’m very grateful for, it stoked my curiosity and, you know, curiosity had been sort of sitting there at the top of the building blocks, and, you know, and I’m like, curiosity is back, and this is exciting. And so who knows where we’re gonna go with this? The only thing we know is we’re gonna go for one more show. It’ll be our season closer this year where we’re gonna take all the stuff we’ve just processed and see if we can distill it into some, you know, big headlines, big takeaways, you know, and.

Michael Horn: Wish us luck.

Diane Tavenner: Yeah, exactly. Exactly. Before we wrap, what have you been reading? Listening to, watching.

AI Amplifying Essential Skills

Michael Horn: Oh, can I do reading? I polished off Stephen Kosslyn’s “Learning to Flourish in the Age of AI.” So it’s relevant. Talks about how AI can, in effect, be a cognitive amplifier loop, he calls it, to the skills that are still important at a headline level. You know, critical thinking, communication, emotional intelligence. He puts in there. And then Angela Jackson’s “The Win Win Workplace.” So those are my two that I have finished.

Diane Tavenner: So we’re sort of falling into our oldest patterns where you’re reading really smart and intelligent books. And I’m blowing through Madeline Miller as I read “Circe.” And now I’m doing “The Song of Achilles” in our run up to Greece. And here’s what I will say. Here’s the connection of “Circe.”

I mean, I just thought it was such an interesting, beautiful book about a female coming into herself and her identity and identity development, as a young woman and then a mother. And it’s just fun and fast, and I enjoyed it.

Michael Horn: That is awesome. Love ending it there. And, you know, look, if AI is really efficient, we’ll have more time to do the reading around humanity that we should be doing all along. So let’s leave it there. Can’t wait to be in person with you for our final episode of the season. And we missed a bunch. We know it. Send us all your hate mail so we can get smarter.

We appreciate you all, and we’ll see you next time on Class Disrupted.

Get stories like these delivered straight to your inbox. Sign up for The 74 Newsletter