Chong-Hao Fu on Putting Teachers Back at the Center of AI

What's a better investment: a data center full of chips that expire in two years, or the people whose learning compounds across a lifetime? Chong-Hao Fu thinks the field has the answer backwards.

Chong-Hao is CEO of Leading Educators, the national nonprofit behind some of the strongest evidence we have on what actually moves the needle for kids; including being the first professional learning organization in a decade to reach Tier 1 What Works Clearinghouse status. He sat down with Alex to work through what responsible AI in schools really demands, and why the teacher keeps getting left out of the conversation.

Learn more at leadingeducators.org.

aiEDU: The AI Education Project

• aiEDU.org
• linkedin.com/company/aiedu/

Transcript

Investing In Humans Over Hardware

SPEAKER_01 0:01

This hundreds of billions of dollars we've been saying spending on AI data centers, most of the chips that are going to these data centers are going to be out of date in like two years, three years. And if you instead invested those hundreds of billions of dollars in humans, humans don't have the same shelf life over generations. Learning kind of compounds.

Alex Kotran 0:22

So Chang Hai Fu is CEO of Leading Educators. And if you are a teacher or if you're an administrator, you probably know leading educators. And if you're not, they're a national nonprofit and they partner with uh education systems, really school leaders, sort of like the maybe the full stack of leadership development and enablement, coaching, in-depth training, um, curriculum, and I think a fair bit of strategic sort of support for schools, right?

SPEAKER_00 0:52

Yes.

SPEAKER_01 0:52

All is accurate. Yeah, that's that's great. You're very close. Yeah. So the way that we think about it is that there's two parts to our work. One is around scaling great teaching, and that's based like everything you said, around coaching, around district advising, around making sure they're the right conditions for adults to do their best work for kids. And then the other part of our organization, uh, which brings us into some of the overlapping work we've been doing, is really around future models and how we create the models for thriving in the future, particularly around AI-enabled instruction, into a changing teaching role. And we and we think both are necessary. In part, I think about my nieces and nephews. They're all nine years old, all in the third grade. And if I think about their experience, I want them to have a really amazing third grade experience. But the real test of any of this is that they thrive as adults. And so we always have to kind of work across multiple timelines in education and probably now more than ever. So that's how our team is set up.

Alex Kotran 1:46

Yeah. So, and you were y'all were doing this work before AI came onto the scene. What or before ChatGPT came out of the scene? Before ChatGPT, yeah. To be stuck, yeah. Yeah, yeah.

SPEAKER_01 1:58

Um the listeners know. I'm

What Leading Educators Actually Does

SPEAKER_01 2:00

not, I'm not, that's not new reframing here.

Alex Kotran 2:02

I think in education, there was for a lot of folks this you know, oh crap moment of this technology that had previously been sort of you know lived in sort of the innuendo in conversations about the future of work or the fourth industrial revolution, and suddenly it was in classrooms complicating teachers' lives in the form of you know students using it as a shortcut or as a cheating tool. How was your org sort of first starting to kind of get the incoming? Was it your partners coming to you and saying, like, hey, what do we do with this thing?

SPEAKER_01 2:33

Yeah. So it's it's really an interesting situation where we got a McKenzie Scott grant almost four, four and a half years ago. And we decided that we were going to use all the funds for innovation because we realized there's a set of things that district partners were asking us to work on together. And then there's a set of things that no one is gonna ask us to work on yet because that just wasn't a burning fire, and there were other priorities that were front and center. And from that innovation team, we actually got connected to the Google for education team. And this was just after COVID. And their team was really trying to better understand the value add of technology in general. And so we worked with them on a project uh to try to codify what are the value add of technology in terms of efficiency plays, in terms of deepening learning, and then also what are some of the transformational opportunities that might exist, both for ed tech in general, but then AI specifically. And that actually kind of formed the basis for us as a team that, oh, we should dig into this because we're not really an ed tech company to begin with. We are around, we've been around for 15 years, but we're really, as you said, known for a as a professional learning partner to school systems and a design partner to schools. And when we started digging into technology, we realized, oh, there's 9,000 ed tech tools. And oh, there's actually not a lot of data that supports what strong implementation at particularly at scale looks like. And so that raised a lot of questions for us like what would it actually take for AI and ed tech in general to meaningfully change learning, uh, to deepen learning, to make teaching more efficient, to then open up new transformational opportunities. Those are kind of the things that we started grappling with. And from that, our first partnerships were actually directly with schools, because we've we we've known over the years that when principals and teacher leaders are working together, that's actually when we see real change for kids. So we just launched a school teams collaborative with schools across the country who are really eager to work on this. And we started the question like, how do you actually want teaching and learning to be different? Not what AI tool adoption looks like, but what is your vision for learning in this new era? And then from that, try to work backwards to then think about what tool integration and then teaching would look like.

Alex Kotran 4:36

Yeah, I mean, we're I think we're such kindred spirits in the in the focus on like the human side of the equation. Well, because you have this really important role. I mean, the fact that you don't have a tool, because I feel like we're in the same boat. Although we are starting to build some tools. Oh, we I I definitely want to dig into that. Yeah. Um, because we are as well. And we're starting to we're having our own kind of like identity crisis of like, wait a second, we we used to talk about ourselves as not a tool developer. Um our interactions with school districts, as I'm sure yours are, part of the reason that they they trust us to give them the advice is we're they know that we're not going in with this implementation outcome as the goal. That's right. And you said something really important to me, which was like, what is the what is the AI in service of? Um how I mean, how do you feel about I mean is AI are are you more excited, more

Vision First AI Adoption After ChatGPT

Alex Kotran 5:30

nervous? Um now that you've had time to process.

SPEAKER_01 5:33

I think both. I think you I think there's a there's a part of me that's incredibly optimistic, and there's a part of me that is terrified, and that is actually what keeps me going. There's a famous Hunter S. Thompson quote from the Rum Diary, so I'm gonna get wrong, but he talks about there's these two poles, like wonderful optimism and relentless despair, and it's the tension between the two that keeps them going. And that's often how I feel right now. I feel like we as a society and as an education system, we're at such an inflection point, and depending on the choices that we make, it could go either way. Um, I think it's actually really revealing that in the United States, we have the least trust for AI. And I actually think that's in part because we don't have a lot of guardrails and a lot of guidance around what responsible ethical use looks like. I think the latest stat is there are 28 states that have AI guidance, but only three states where that's actually enforceable. The rest is just kind of like it'd be nice to. And, you know, as you know, the models that we have are benchmarked around speed, around accuracy. They're not benchmarked on learning or human development or relationships. So some of the things that we're optimizing for are not necessarily the same things that we care about in the education system. At the same time, we also know the job of the teacher is incredibly unsustainable. I think the latest stat that we have is only 19% of teachers think their job is doable. And we also know that AI can be huge in terms of time saving, in terms of ensuring that kids get the personalization that they need. But all this has to be done, I think, really responsibly and carefully. And so one of our mantras throughout all this work is how do we learn and not guess? How do we test alongside educators? How do we actually talk to kids? How do we know how it's going? And like, let's just not assume pro or con. I think Aaron Mote talks about AI as an arrival technology, like electricity. Like, I think I'm sure you've think similarly, like electricity is neither good or bad, right? A lot of it depends on how we use it and what are the guardrails and guidelines and productive use cases that we all discover together.

SPEAKER_03 7:32

Yeah.

Alex Kotran 7:33

The because it's really complicated. Yeah. Also that there's I mean, even if the even even because I agree that most AI models are not tuned for learning. I think that there's actually some like we're starting to see some tools and use cases that are that are really exciting. And I think there are organizations, some of them nonprofits, that have like started to dial in what what does it look like to actually ground the right context and have like the right you know pedagogical frameworks, you know, high quality instructional materials as the reference point, you know, for tools. And that wasn't always necessarily the case, but it's more and more accessible now. Um but that's still sort of there's this bigger question, which is like, what is the point of the learning? And it and if and if the tools are really good at increasing, let's say, scores on the existing assessments, I think that's happening in a moment where there's a lot of questions about, well, are we even assessing the right things? Um how do you approach those conversations? Because we we we get asked a lot of these questions about well, like, well, what about are we even measuring the right the right stuff? And I think what's hard if you're a school leader right now is you don't have a ton of agency over that. I mean, you can you can go and advocate and you can sort of put your voice out, but at the end of the day, superintendents don't get to decide what the assessments are. They sometimes have limited ability to actually dictate, you know, the degree to which the assessments should influence, let's say, you know, um like teacher incentives, uh, like even compensation. Why specific like even specific to compensation? Yeah. There are states where teachers are literally paid based on how students are performing at these tests. Um how do we deal with that that aspect of like, yeah, there's like there's a lot of stuff in the system that needs to change that you don't have power over.

SPEAKER_00 9:25

Yeah.

Alex Kotran 9:26

But I think people need to feel agency to really to not be sort of sucked into the feeling like victims, I guess.

SPEAKER_01 9:32

Yeah. I I love this question. And I love the report that you all put out with with Burning Glass around how student learning outcomes need to change in this era. I thought it was really insightful. And I really appreciate the point that the report made that it's not just about what content we assess, but actually how students learn. And one of the ways that we think about it is that there's a backwards design problem right now. That it's like we can't think of it as like a teacher burnout problem and then a technology problem and then as an assessment problem. It's actually all one problem. And the problem goes like this we need to re-examine what's most important for students to learn and master. We then need to go backwards and then re-engineer what are the most important learning experiences that are aligned to those outcomes we want for students. And then from a resource layer, we need to think about well, how do we create those learning experiences through a combination of really amazing teaching, but also technology that is reliable, ethical, and tested. And so if you actually backwards design from what we want from students, we can maybe tackle multiple problems at the same time. The problem, like you're saying, is like there's not a lot of time to do that type of thoughtful backwards design. We're seeing a lot of when we think about AI adoption, uh, our VAT framework that we developed talks about how the value add for technology is often about what teachers can do more of, what they can do better, and what they can do new. The do better part is around efficiency, and that part tends to be really easy to do. People can think about new workflows, they can think about lesson planning differently. That happens pretty much on its own. The do better to actually deepen learning, we find that has to be really carefully tested because if not, you could actually create a situation where introducing the technology results in cognitive offloading and less, less learning. And then finally, the do differently, do new. That that is the part that I think people um need the most time and the most support in processing. Well, what is the new vision and what's most important in this era? And I thought like the report that you all put out with Burning Glass is like really, really helpful. We're trying to think of ways in which you can integrate the most important skills for this new era inside existing content disciplines. So I'll give you an example. We're we're engaged in an effort right now to think about what it would look like to modernize mathematics. A lot of that is actually about how we learn math. Within the math standards, we have the content standards. We also have something called the standards for mathematical practice. And that's really how mathematicians think and do math. It includes like argumentation, precise thinking, using tools, modeling. Those are maybe the more evergreen durable skills that would transcend outside of mathematics and be really valuable in the age of AI. It also happened to be the skills that in a math classroom get deprioritized because they're not assessed in the same way. So, how could we really think about how we learn math as a vehicle to learn things that will be really durable and really cultivate a lot of agency and a lot of transference in this new era? That's like one project we're actively thinking about. And I think this is could be this golden age of teaching where we create these moments where students are more agents of their learning, they have a lot more drive and purpose, and that and it requires

Trust, Guardrails, And Teacher Burnout

SPEAKER_01 12:30

a stepping back. And I forgot who said it, but I love this quote of like, this is the time to marie kondo our learning standards and maybe kind of think through what is no longer serving us and be really just create space for what what future opportunities might emerge.

Alex Kotran 12:45

Yeah, and what you're describing is not necessarily well, we don't need to teach math anymore because the AI can do math, or we don't need to teach you know, writing skills anymore because the AI is good at writing. It's it's this intuition that the domain knowledge by itself, which I think has been the primary goal of a lot of core subjects, because that's what's tested. That isn't necessarily the most important endpoint. It's sort of like uh almost a side effect of really good math instruction or English instruction, but actually it's more like do you have the ability to you know form your thoughts into a coherent and cogent argument and then you know make the case for a position that you've established. Um and whether you do that with a five-paragraph essay or with you know some different modality, that piece may be less important. But it's not so much that you don't need to be able to think.

SPEAKER_01 13:38

I think that's exactly right. And the disciplines that we've created have a lot of opportunities to think, to have like real knowledge of the world. And the deeper and denser that knowledge can be, I would I believe that that allows students to then acquire and learn new knowledge. And like being able to go really deeply into dinosaurs gets you excited about learning perhaps about the next thing that you're going to dig into as a student. And it isn't necessarily that we're deepening these skills just so that you can only do those skills. It's also so you can have discernment and judgment and application.

Alex Kotran 14:10

Yeah, I had this interaction with a student who she was in one of our vibe coding workshops that built this amazing tool. And she'd never she her school had banned AI. So it was it was interesting to see somebody absolutely running with it, like outrunning, outpacing adults. This was at South by Southwest. It was like a tech conference. So these are like tech forward educators and ed tech leaders. And you had this 17-year-old who never used AI before, but she had domain expertise in um uh non-pharmaceutical therapies for ADHD. Because in she had a class, literally, the class that the class was pick a subject that's interesting or important to you, and then do a full semester, maybe even a full year, of like in-depth research, like rigorous academic research into the topic. And so she had just finished that class. Do you remember what school this was at? This is one of the best schools in the country. This was like a private school in Connecticut. Okay. Um so that specific the the the school itself is is not necessarily something that every every district can necessarily carbon copy. Yeah. Um and that modality of like having a dedicated class, just like they're whatever you want. I I know that for some districts that's probably not within the realm of possibility, but maybe it is. I mean, yeah.

SPEAKER_01 15:28

I mean, I feel like I've seen versions of that class. Sorry, I I don't actually know exactly what the class is, but I've seen versions of like a much more student agent, student purpose-driven high school class in several places across the country now this year. And one of the things to your point that I keep hearing is it's much, much more common in high-income schools. And I think we, I think what the latest stat is when you're looking not at just AI replacing teaching or a kind of AI towards existing educational outcomes, and you're actually looking at AI fluency. I think the current stat is 66% of high-income schools are offering that to students and only 6% of low-income schools. So an 11x difference, which obviously is a huge digital inequity. One version of what you just described that I saw uh recently. Sorry, did I interrupt your story? No, no, no. Okay, I feel like I Okay, we're okay, uh, I apologize. Um at the Austin and Richards School for Young Women, they also do a senior interdictionary project. And I was talking to two seniors who whose moms had both suffered for seven years from uterine fibroids and who should have been offered a hysterectomy much earlier in their healthcare regimen, but that's often underdiagnosed in communities of color. And so both high school seniors, to your exact point, were passionate about this topic. They research the medical treatments, they met they research the lifestyle factors, the demographic factors that influence uh needs for hysterectomies, and they created an app on Play Lab that then can coach you in any language, if you're a woman struggling with uterine fibroids, to then ask the right questions of your doctor. And it's like that example of when kids are really proximate to something that they care about, they're given the space to explore it with a lot of agency. And then AI actually becomes almost like a magic wand, if you will, to create something really meaningful and authentic that they can offer back to the world. And that I've seen so many examples of it this year. It gets really, really exciting. And then we also, and you juxtapose that against the fact that I think it's only 20% of high schoolers find high school to be really relevant and engaging to their dreams.

Alex Kotran 17:32

And that's that low. I knew it was low.

SPEAKER_01 17:34

I didn't realize it was 20%. It's really low. And so then the question is like, how might we create more spaces in the existing curriculum for kids to examine their purpose and passions? Because my my sense is just creating a like it doesn't have to be the entire day, but creating those authentic experiences are gonna create the inspiration and the motivation to connect all the rest of the learning back to what they want to do in their future.

Alex Kotran 17:58

But this is actually so this is where I was going with uh you you took us there. So it's A was really cool to see that, like, yes, almost like duh, if you give kids, if you have kids the space um to take agency agency with their learning, they're gonna do some amazing stuff and it's gonna be connected to what's important to them. But my head also goes to most of the schools that we work with don't have the luxury of every kid having that dedicated time. Um, because if your math scores and your reading and writing scores are you know below where you need them to be, feels like all of the time that we have should be focused on, you know, addressing that gap. And and you mentioned something about how do we bring this into the core subjects. And so this is like this, it wasn't even a debate, it was more of a a curiosity.

SPEAKER_01 18:44

Yeah, go ahead. You have a gap-filling thought. I mean, I think there is this sense that we have to only target the gap, right? And it does actually become a little, it can feel really painful in that way. One of my earlier experiences, I back in 2006, 2007, had the opportunity to open KIPP's first fine arts campus. And I taught at the original KIPP school and then realized that I wanted to think about how we might create more authentic expression and more agency for students. And this was through the fine arts. So taking a taking a left turn from AI for a second. But all of our kids took dance, theater, orchestra, choir, visual arts, and so they could choose the fine art form that they wanted to major in over four years with us. And one of the things that we would see over and over again is you would see a student maybe struggling in English or math class, but then they would get the solo in orchestra or they'd get the lead role in the musical. And suddenly that sense of pride and that sense of confidence flooded everything else. And they actually went back and did so much better in, and because there was a new reserve of confidence that they could pull from. And I think sometimes we take like a only a weakness-based approach when really in life, we're often measured by our strengths and what we can do the best. That's what we we ground our professions and our personal lives around. And doing that for kids also can be really, really powerful because then you're not only practicing your worst thing, you're actually building reserves of confidence and energy and relationships that you can then leverage into weaker areas as well. I found that over and over again. And so it just thinks about like where can we create these spark magic moments? And I I do actually think AI fluency, AI creation, vibe coding can create really amazing things uh really quickly.

Backward Design And The VAT Framework

Alex Kotran 20:30

Maybe the obvious thing that I that comes to mind is okay, so you have like let's say history class, and I studied history, basically political science and history in college. Um and it I'm just imagining if if if I if history class had had a whole semester where I got to be, you know, pick a pick a pick a time period and just go really deep and become an expert in that time period.

SPEAKER_00 20:54

Yeah.

Alex Kotran 20:55

Um, you know, I don't really use any of the knowledge that I I don't think I use any of the knowledge that I got in history. So on the one hand, that was sort of a that was a a thought experiment that I put out, and I forget who who I was sort of sharing this with. And their response was but there's actually something really important about having sort of like a broad knowledge about sort of history and how we got to where we are. And so is it okay to like miss out on that? Like, and I and I guess that brings us to if you're gonna start to get into core subjects, how do we determine, you know, what aspects of let's say math or or English or social studies or science are the non-negotiables? And then where do we start to identify the places where it's like, oh, is this actually critical to being a citizen? Um, I think AP European history probably is an example of something that I could have swapped out with maybe a slightly more open ended uh course.

SPEAKER_01 21:48

But do you really believe that you don't pull from your historical knowledge in understanding the world today?

Alex Kotran 21:54

I believe I do. Okay, you do. Yes. Okay. But I think it's if you were to ask me like what asp what like What history class or what semester of history class could you have gotten away with? Like not having gone through? And I think it's hard because you'd say European is what I'm hearing. Yeah, but that was that's because it was like I'd already sort of completed all of my sort of like requirements. That was like a senior year kind of thing.

SPEAKER_01 22:16

So I I have some theories here, but I think this is a place that we do need to do a lot of further study together as a field and to have some honest conversations around how we might Marie Kondo and how we might preserve the standards that are most important. Um well, first thing is we know, particularly from reading, that knowledge of the world predicts your ability to read all kinds of things. And it's a density of knowledge, the vocabulary map that exists in your brain that allows you to read a wide range of texts. And actually, what makes someone able to read a text like Immanuel Kant in college is less about like reading ability and actually about knowledge of the world and all the interrelated content that's there. And so there is this danger where if we cut off knowledge, bodies of knowledge of the world, we do actually make further learning on those related tracks more difficult. But we also know obviously we we need to get uh we need to kind of focus a little more. So in my mind, the visual that makes sense is one of like a tree where there is like a tree trunk of core knowledge that we might expect all students to really deeply understand and master that goes maybe up to sixth grade. And then very soon, then there a branching begins and then different pathways emerge. And the smarter that we can be about durable skills and like what are required for different professions, and the more honest we can be with students, we can then say, like, hey, going down these knowledge pathways open up these doors, but then they might close others. And the earlier that we can help students be agents of the pathways that they choose, the knowledge that they acquire, the skills that they acquire, you start to maybe imagine a different learning system. I think Diane Tavner uh at FutureMe is doing something similar where she's trying to map out all of the different professions and then all of the skills progressions that would lead to different possible learning places. Um, I think some of the curriculum companies are also trying to think about how might we take our high-quality instructional material. And instead of having all students progress in a linear way, what might a kind of nonlinear progression through our curriculum look like? And what are the different pathways? We know there's not just one way through the math standards. There are probably multiple ways through. And those pathways depend. Like if you are planning to go on a calculus route, if you're trying to go a data science route, if you're trying to go like a general math route, there might be different pathways through math where all of them set you up with the ability to think critically, to analyze data, to make arguments, to model, but they might also point you towards different career pathways so that we don't have to kind of jam the curriculum full of everything, but students are still having a lot of agency and choice around the paths that they want to progress. Potentially lots of equity issues there, potentially lots of guardrails and safeguards that need to be put in place. But theoretically, we could be helping students make more informed choices around narrower standards earlier in their high school in their secondary progressions in particular.

Alex Kotran 25:11

It feels like it's gonna be really hard to get this right without being in a place where we can really trust teachers to do this and make some of the calls, especially in this exploratory moment where you know, in 10 years maybe we will have enough data. I I'm I'm incredulous of that, but let's just go with it. Let's say in 10 years we do everything right, we have all the research, and we're like, okay, here are the new standards, whatever they are. But in that, in that journey, there's gonna be a lot of I think experimentation, and we have to be careful, right, that we don't condemn an entire generation of students to potentially bad outcomes, you know, while we wait for the right state. I think that's where I'm I'm caught between, yes, I knew I think we do need more research. I just I worry about going to a school and saying, well, you should just sit tight and hold off on innovation until we have all the answers. Maybe we won't have any, and we won't never have the answers by the time we have all the answers from the research that we're doing now. I mean, the career pathways might shift again. Are there any examples that you've seen where schools have done a good job? And let's focus maybe on underserved communities so that it's you know, schools working within like like resource constraints or constraints of of other kinds that have been successful in changing the way the teachers see themselves. You mentioned like 19%. I think it was is that the same stats like how many teachers would recommend the profession? I think there's something similar to that where it's like a vanishingly small portion of teachers would recommend the profession to somebody. It's 14%, I believe.

SPEAKER_01 26:48

14% that would recommend the profession. For the friends or family, I think the net promoter score for being a teacher right now is around negative 50. Wow. So two things here. One, we really believe working alongside educators and co-creating is really, really important right now. We know that in times of uncertainty where conditions are shifting, you can't just be some expert in an ivory tower and just assume that that's gonna work. You actually need to be designing alongside educators, helping them understand how to learn responsibly with students. Like that's where our work is. And when we have gathered teams of school teams together or district teams together to work on how they're gonna evolve teaching and learning in the age of AI, educators get really, really excited. And it's a conversation where you can be real about the risks, but also really, really concrete about the visions that we want for kids. Because right now, almost every district has a portrait of a graduate that they're hoping all of their high schoolers graduate with, but it's not always concrete and it's not always tied to changing conditions in terms of tech and society. I think educators really want to be a part of that conversation and to shape what that looks like and to kind of reimagine how classrooms and school spaces and um community collaborations would look. And folks get really excited. You know, our teams ran together this rural AI Schools of Excellence collaborative with districts across the southwest. And I was pulled aside by a superintendent and teacher recently, being like, it was the most impactful professional learning they'd ever done. And it was really because they were thinking about well, how do we make the vision that we want for all our learners real in our schools? And just have the time to collaborate and think that through with expertise, with insights. The other thing I would just say, in terms of leaving a generation behind, uh, in terms of what content matters, I think there's a both and here, back to the math example we were talking about earlier. You can learn both the math content and how you learn it could incorporate the durable skills so that it is kind of more interactive. It is more around the modeling and argumentation. And I hear a lot of times it's either one or the other. I was recently talking to a superintendent where they were really excited that um students were really collaborating and really discussing. And at the same time, one of the concerns we had is the collaboration was happening with a text that was, you know, these were fourth graders using a first grade text. And so that's, I think that's a false trade-off. We don't want folks to be collaborating, but on something that's wildly inappropriate in terms of what the science of reading says. We have to kind of hold both what we know about the content disciplines and evidence base there, and also make sure that we're um thinking about how that lines up with durable skills. And I think that's going to be really challenging. And so one of the folks we just brought on our team, she actually does came from an MIT lab on play-based assessments. And so her whole lab was like specializing in like how do we actually help students reflect on their metacognition to reflect on their agency, on their problem solving while they're in the content disciplines. And I think we're experimenting now, like, how can we make that really easy for educators to deploy in their classes or for kids to use? Because like we're talking about adding another layer of things on top of the content. And that that could be really challenging. But that might also be the place where technology or other things open up. I think we were both in London with the Google DeepMind team. I don't know. Did you play if they're new durable skills assessments? Yeah. So like it's kind of, I mean, they're, you know, they're they're still working on it, but there's like ways to then assess collaboration and I think conflict mediation they had. But like, I think we I think there's going to be a whole nother set of things that we pay attention to, but we have to find ways to do it such that we're not overwhelming students and educators. There's a lot of ones.

Alex Kotran 30:29

No, I'll take that. I I think so so the answer might not be you throw out an entire semester of social studies. It might be that, you know, can you identify research projects that are aligned to the standards that you're supposed to be teaching, um, but that give students a chance to go and you know learn about history in the way that it intersected maybe with their community, right? And like giving them a project where they're going and interviewing if I was a history teacher, I would have kids go to a um uh senior living homes and interview residents and sort of like collect these sort of like oral histories of different, you know, historical topics. So so yeah, I I I I I I like this idea that there's not we like literally shouldn't have to choose between the two. That's sort of the point that I'm hearing.

SPEAKER_01 31:15

And it's often built into our standards. We just are not kind of pulling that part up. So in the science standards, to your point about like learn about what's happening in your community, we know that phenomenon-based science is a core part of the new science standards, which is literally

Student Agency Projects And AI Equity

SPEAKER_01 31:30

helping students understand the phenomenon that are happening in their communities. The best science curriculum and the best science teachers are already doing that. It's almost kind of pulling up how do we lift up that lens of inquiry and curiosity exploration such that more students are getting access to that type of instruction.

Alex Kotran 31:48

That was one of the criticisms that I got after we published the the Burning Glass report was you looked at standards, but the standards actually are not the same as the assessments. It isn't the case that schools are assessing all of the standards. And one thing that you you you said is you know, AI could help us with that assessment. I just want to just you know, for our audience, like if you're not a education policy nerd, like I think part of what you're getting at, if I may just draw it out, is we have all these standards for durable skills. If you actually go through this through uh core subjects, you'll find critical thinking, for example, across I think all the core subjects, even in computer science increasingly. Um but in the in the tests that students take, they we aren't necessarily evaluating them on critical thinking to the extent that maybe we should. And uh part of the reason is because like there was this test in uh uh Jeff Riley, who's the former chancellor of education for um state of Massachusetts, was telling me about when he was at Baltimore Public Schools. We heard the Ms. PAP. I don't know the MISPAP. Um this was this was uh one of the most famous standardized tests uh because it was pretty much the only standardized test uh that uh in Jeff's recollection that his students were excited to take because it measured uh uh uh collaboration, communication, durable skills, critical thinking. And you student basically students were on teams and they would go and basically work on problems together in groups. And the test was basically a casualty of No Child Left Behind because students weren't given an individual score, they were given a score at the building level. And in sort of this moment where everybody was obsessed with having like really, you know, like specific data to each student, it was sort of seen as like not the right use of time and resources. But Jeff's point was like we know how to do this, like it just part of the challenge is like logistics, you know, it's sort of actually just really hard to do things like measuring, you know, students, you know, assessments that aren't on a Scantron. Um, it becomes a lot more qualitative. And how do we do that in a way that's you know fair across multiple, you know, teachers, let's say. And I think this is perhaps where AI does offer some opportunities. And it's like I have to be careful because I'm I'm worried about relying on the AI for assessments. At the same time, it does feel like it could be key to unlocking, you know, some of these different ways of measuring. And also like if you talk about you also talked about uh iteratively adjusting the learning path based on what how students are sort of learning and or how they're performing on, let's say, sort of like interim assessments. And I feel like that level of personalization does require some additional tools. How do you do you think this is the moment where we should be like leaning into like with respon like with responsible uh caution and oversight? But um like has your team started to actually work with schools on piloting some of this stuff?

SPEAKER_00 34:45

Yeah, I mean, I think you said so much. Um, one, I love the idea of it.

SPEAKER_01 34:51

If you're free to respond to any one of those, and well, I love the idea of a collaborative assessment that gets at at things that become hard to measure be as a as a sign of what we what we care most about. And I I wholeheartedly agree that assessment and formative assessment and a whole new type of assessment, ambient assessment, automatic speech speech recognition. I think there's a whole world that's opening up in terms of how we can lift up data for students in terms of their own learning and also obviously for educators as well. I think there's some really powerful things that are emerging right now. And things like obviously like Kidum and Atlas, so it's where it's built even right now, such that formative assessment on today can inform what the lesson looks like tomorrow. I think that there are many folks who are building things in that space. And I think I would follow for sure to your point about caution, I would definitely follow the evidence base. I would look at how those tools are being tested, whether it's an A-B test or something more rigorous to understand does it actually improve learning, does it improve agency? I would obviously want to be really cautious about how the tool is designed and how those assessments serve a wide and diverse range of students, and obviously how then the data is used and who owns that data over time. But with those caveats, it seems like a very, very powerful feature because I think what AI will be very good at, obviously, is to analyze the data, surface the insights, and then to make personalization happen. And that moves the human, the educator in the room towards a more relational, more inspirational, more social type of facilitator of learning. And I think that is almost certainly the direction we're gonna go, assuming that we do these things responsibly and we actually earn the trust of educators, of students, of community members, because I think the other risk of all of this, as we're seeing obviously in some states, is that if it's not done with a lot of transparency and a lot of evidence base, we're gonna see all kinds of pushback on technology. And I think that that's I think we're gonna have some really uh I think we are kind of at an inflection point.

Alex Kotran 36:53

Yeah. Are you you are you encountering a lot of pushback? We definitely are.

SPEAKER_01 36:57

Um not yet. But I think our our stance is really around helping districts use the tools responsibly. And so I think there's less less pushback on us directly since to your earlier point, we're not necessarily a tool provider.

Alex Kotran 37:12

Yeah, we I right. We're we're not encountering pushback on AIDU. We are we are our partners are encountering a lot of pushback. Oh yeah.

SPEAKER_01 37:19

And so wait till the new Toy Story movie comes out. Oh, is it AI? Yeah, the the villain is an iPad and AI.

Alex Kotran 37:27

Oh my god. What is your stance on screens? And I won't we don't have to talk about screens.

SPEAKER_01 37:34

Um I think it's about discernment. I mean it's about discernment, right? And making sure that things are age appropriate and high quality and and useful.

unknown 37:41

Yeah.

SPEAKER_01 37:42

We can talk about screens if you want.

SPEAKER_00 37:44

I'm sure that I'm sure you've got a lot of guests. I don't know.

Alex Kotran 37:46

I don't I don't know what my position is on screens. I mean, I I roughly do. I mean, like I I think it is probably a good thing for kids to spend less time on social media, broadly speaking. Yep. And to the extent that that is in interfering with their learning, definitely finding ways to address that. Um at the same time, it's hard to imagine how you could do a let's say in-depth research project that makes use of all the amazing research tools at your disposal if you're don't have access to a Chromebook and or a computer of some kind. And so are we robbing kids of are we literally throwing the baby out of the bathwater in an effort to protect kids by actually like removing something really important uh you know learning tools that you know whether they're necessary or but certainly extremely valuable um in certain in certain learning uh uh contexts. But that resonates? Yeah, yeah. But I don't know how to I don't know how I would respond if I was a if I was a superintendent of a school where you know the parents are you know showing up to the school board meeting and if they're like virulently anti screens, is that the hill that you die on? I'm not sure. I one thing that you you talk about is what before we before we move too far from this topic, is I want to hear more about your team, like like your experience building some stuff.

SPEAKER_01 39:06

Okay.

Alex Kotran 39:07

Um, but you did uh what was the language that you used that I wrote down? It was um Have you been writing?

SPEAKER_01 39:15

I'm I'm impressed that you're able to No, no, this is just like my pre-notes.

Alex Kotran 39:18

And I sort of just like I had the AI try to like simplify it a little bit. It wasn't it's actually now I'm just like reading the AI synthesis of what I had written down. So um AI has not been helpful in this situation. Um okay, you you had this concern about cognitive offloading. Absolutely. And usually this is a conversation that we have when we talk about kids. And, you know, is AI going to get in the way of their learning? And we can have that conversation if you want, but I'm actually really interested in, you know, some of the ways, some of the times that I've heard you talk about this, it's actually in the context of teachers. Yeah. When you talk about efficiency, we hear a lot about, oh, teachers could use the, you know, use AI to, you know, shorten or differentiate their lesson plans. How like if you're a school leader, how do you, how do you advise a school leader who's coming to you and saying, well, you know, I'm thinking five, 10 years down the road. And if all of our teachers, if their instinct is just to rely on the AI to do the hard stuff of like actually taking a lesson and like actually redesigning it to be more project-based and career connected, like that actually is a muscle I want my teachers to have. Um yeah, I mean, yeah. How do you how do you respond to that?

SPEAKER_01 40:26

So we're so we're doing a lot of this right now, and we're working with four different districts to design uh AI-enabled tools linked to illustrative mathematics to help teachers internalize the content, rehearse and practice the content with each other, and then actually scaffold the content based on where their students are. And so we've gone through three cycles of data, and it's I think it's kind of fascinating. So, and I think there's a lot of lessons for us as leading educators and some humble lessons that we should be taking in. So I don't think these are actually lessons necessarily about teachers and learning. It's maybe us and kind of us learning how to be better tool designers as well. And we started trying to build tools based on what we know about good instructional coaching, which is that you always want the coache doing most of the thinking. Right. And so as we as we plan for any lesson, like you want to be thinking about, well, what's the most important math in this upcoming lesson? What does what does mastery look like by the end? And what would I want to see in the exit ticket to indicate that a student really gets this math concept? And what might be some of the most common misconceptions

Rethinking Core Knowledge And Pathways

SPEAKER_01 41:30

or missing prerequisite skills that will surface? And how do I then, through questioning and examples, scaffold for those missing pieces? And so we basically designed um internalization protocols where teachers, if you had like five minutes, if you had 10 minutes, if you had 20 minutes, EA would actually kind of quiz you on your upcoming lesson. And it would be like, hey, okay, what's the most important math? You have the student Alex, based on their data from the last unit. Well, what might you expect they might struggle with in this next lesson? When Alex says that, what question might you ask to respond? And because then the AI coach had then been trained on illustrative mathematics and also the knowledge graph of math standards, you would expect that it would give really good responses. Well, what's interesting, like, so that's an app that doesn't give you a lot of um answers, it gives you a lot of questions and forces you to do a lot of thinking. And so we did actually find that usage was less than we expected, the more that we didn't provide answers. And so we're trying to balance like, well, what's the right amount of sweetener that we are like answer giving that we want to give as work is happening? And we also saw predictably that depending on how experienced the teacher was of the curriculum, they were then less likely to use it. But we actually have also seen, interestingly, teachers who use the questioning app uh pretty consistently are then more likely to actually use illustrative mathematics and teach the hardest parts of the lesson and teach the lessons with a higher degree of skill. So that's kind of encouraging. But then total implementation rate, still only 10%. So that's not great. And so then now we're thinking about well, what is the actual social context? And when would you want to be using this tool, whether it's in a professional learning community or in a coaching visit? Like, what would what would this tool actually replace and make easier for teachers? Because we can't just design something that adds more work, even if we think that it's helpful. And so we're really trying to think about how do you design it so that it fits into the natural workflow of an educator, so it doesn't feel like an additional thing to do, but also creates a lot of thinking and not just like answer giving. Yeah.

Alex Kotran 43:27

So then no, no, it totally makes sense because ultimately they can just go to Chat GPT if they're like, for sure. Like, no, I just want the lesson to that's right. I don't want to have a Socratic seminar right now.

SPEAKER_01 43:38

Don't patronize me, right? And so I think there's also a way to kind of like, are you in the mood for a Socratic seminar? Are you in a mood for just a direct answer? And then there's some variation. I think we're trying to think about what are the most useful suites of tools for educators. And I think for us, that also means like listening to teachers when they're saying, like, yo, this tool is asking me too many questions and like taking that seriously and not just say assuming that we know better. And that's obviously really, really important. And so that's gets into like user design and testing with teachers. And so these are all new lessons for us right now.

Alex Kotran 44:12

I just I wonder, I mean, do you think there's any I don't know if there's any product based solution to this because I don't think there's do you think there's a I need to learn how to phrase these questions. Do you think there's a product-based solution to the problem of um discernment in the sense of a lot of interacting with AI and AI skills actually comes down to just being self-conscious about, you know what, I could use AI to like to take myself as an example to, you know, write a blog post or you know a Substack post. And yeah, I could probably get away with it, but at certain point I'm going to lose my voice and this will just sort of become the AI's voice. And so I'm just intentional about not over relying on the AI. But I don't know that there's I've tried to create tools that sort of like replicate that, but oh yeah, it either it just doesn't do a good job and I end up reverting back to my previous workflow or or it's it just like it's I find myself just having to like almost like I'm like helping it along. It's not necessarily um like I haven't gotten the I haven't figured out how to build like an AI writing assistant that like sort of like works with me on the the outline. It's like it's often just spitting out an outline and I'm just like working backwards off of like what it thinks that I should be writing. Is that something we just have to come to terms with that like this is just going to be part of being an AI user is getting to a place where we can trust teachers to make those calls. Such a great question.

SPEAKER_01 45:42

I'm I'm not sure is my is my honest answer. I I like the way that you framed it that it is about discernment and you're not feeling like there's anything that actually captures your authentic voice. And then I wonder what are the uncapturable moments of teaching we know from Deborah Ball's work that on average a teacher makes about 200 um moments of decision in a class that there's so many that happen all the time in terms of how you validate a student's response, how you correct an answer that might be slightly wrong, how you get students to work together. There are so many moments of decision. And I think oftentimes when people talk about the the power of AI for learning or AI tutors, I mean it's true that we have powerful tools, but people maybe haven't seen the hundreds of decisions that a classroom teacher is making every day that are about relationships, about identity, about connection. And obviously there are ways to kind of make those decisions better. And I think there's a way in which AI could analyze the data and the choices and actually lift up patterns that maybe are unintentional or unhelpful that actually might yield to the discernment. But at some point there is also that human judgment that kicks in. So I I think about how might AI kind of create the heat map or the dashboard, but then what is the actual protocol that allows a person to change and make adjustments. And I at this moment I think it's interesting where I think some people really like feedback from AI because it feels less judgmental and feel safer than a human coach. But other people obviously still want and value that human relationship to make sure that they are like really seen and really valued and have and the person they're talking to has real expertise to pull from yeah you get to the there's a shortage of instructional coaches.

Alex Kotran 47:30

Yep. And so on the one hand, could AI help to bridge that gap? That sounds exciting. On the other hand, if AI can bridge that gap does that then reduce the urgency for a district leader to backfill the instructional coach position that they've been trying to hire for yeah well one one interesting piece of feedback that we got on some of the tools that we were designing.

SPEAKER_01 47:51

So we're an organization that I think is really good at coaching and professional learning. And one user was why are you designing AI tools around the thing that you're best at the coaching and professional learning why not design the AI for the places where there's actually a lot of drudgery and a lot of paperwork and a lot of data analysis and operations. Help us save those five to 10 hours a week of the things that we don't want to be doing so that we can use more of the time on things that are uniquely human. And that seemed like a really nice place to start. And so maybe if we focus on the things that we think are the highest value add, that are our unique superpowers, where we have the greatest depth of expertise and then start on the places that should be automated because nobody wants to do them and they're not adding a lot of value to learning or relationships, that might be a safer place to start anyway. It feels tied to this idea of you know teachers I mean I think it's amazing.

Alex Kotran 48:46

I think every teacher if they're still in the role right now, I mean they're like truly a superhero because they have had every reason to leave the profession. They're one of the most overeducated relative to what they're getting paid. Like most teachers have more education than I do that I talk to because it's like I don't have a master's degree. And and they're sort of treated as like pawns in this like the grand strategy of education. It's like oh how do we how do we leverage teachers to do this or empower them to do that. But it's um there's something powerful about saying like no you actually have this really important expertise.

SPEAKER_00 49:21

Yeah.

Alex Kotran 49:21

And we want you to like bring that to the table and we yeah like our role is how do we learn from you and yes get other sort of pieces out of the way.

SPEAKER_01 49:31

You probably know have we told you the story that my family is my mom's family is from Singapore. And when I head back to see family sometimes have time to meet with the Ministry of Education over there. And one of the things that they shared is that they have 125 teacher leaders. They're master educators of Singapore. They are paid as much or more than any education official in the country and those 125 teachers are the ones that decide what great teaching looks like at in Singapore. They actually look at all the research coming out of the United States and they decide hey this is what it looks like in Singapore. They write the protocols, they film the videos and then that's what becomes the definition of great teaching. And I haven't spoken to them recently but in my mind at this moment in terms of AI, there's no doubt that those are the ones who are then actually testing and figuring out what it looks like in great classrooms and in great learning environments. Like we that's the power of actually having that lived experience as an educator. Like you can't actually reduce that and there's something really essential and valuable at about at this moment in terms of engaging in terms of policy and what great teaching will look like in the future, which is why we at leading educators are so passionate about co-creating with with amazing educators around the country.

Alex Kotran 50:45

Yeah it feels like the the only way to because the the trouble is we're trying to scale basically change management and like school transformation in service of creating like cultures of innovation and lifelong learning right like it's like it's very it's really hard to scale something like that which is like deeply human. It's complicated it varies from community to community the challenges change uh the resources change um there really isn't a broad asynchronous like learning credential that we just like every teacher needs to get this credential and then they're going to be good to go. But yeah Singapore is an interesting uh example of if you if you really center uh teachers as the main characters in the story this is my new my new tagline is like we we need to be talking about the teachers and then the AI is sort of like the supporting player um I feel like we get to the conversation about teachers through AI and like and I feel like we actually need to reverse that. Is is Singapore similar to Japan and like Korea where it's actually like one of the most sought after like jobs that you can get like you try to get into it is you know become a teacher and if you fail it's like okay am I fallback I'll I guess I'll try to get into medical school.

SPEAKER_01 51:59

Yeah it's really hard. In fact, you know I was talking to the ministry in Singapore they only select 700 teacher candidates every year and they all come from the top of the graduating university classes. And once you're selected as one of those future educators, all of your um undergrad and grad school are paid for and from day one of training you get the curriculum that you'll be using you start to kind of envision what your classroom will look like. All of that happens uh very early in once you're kind of selected as a teacher candidate.

Alex Kotran 52:28

Do you have a sense for I mean

Building Teacher Tools Without Offloading

Alex Kotran 52:30

I I've my guess is that they get paid a lot more than teachers here. It is more um but I think is that is is is compensation actually what's driving that I think it's a respect and the social prestige also is very important.

SPEAKER_01 52:44

Not that the compensation isn't important obviously we should pay our teachers more how how turn to we should pay our teachers more we should turn we should pay our teachers more.

Alex Kotran 52:52

We should pay our teachers more but I also think we should celebrate our teachers I mean I think part of this is what are like the baby steps that we can take to reframing what it means to be a teacher in a moment where we're where we clearly need a lot of help solving this.

SPEAKER_01 53:07

And I I I do really think these are related problems. And so not to think about changing outcomes like changing durable skills that are important for kids, changing learning models and then like the act with technology and then also the teaching profession. The three are connected issues, right? And so if we could engage and lift up examples of educators who are really creating new inspiring learning models, whether they're AI enabled or not and how they point towards a set of really inspiring durable skills, I think those examples are happening all around the country and kind of lifting up it's almost like if there was like a new grand challenge of how AI enabled instruction could really prepare future ready students, I think there could be some really powerful stuff that emerges almost instantly.

Alex Kotran 53:51

Yeah. Yeah and then if you're if you are trying to build a tool that's who do you go to you're going to go to this like set of trailblazing teachers that um are going to be able to give you some really deeply informed feedback about why it may or may not work. Who it's teacher appreciation week that's what is it? It is do you have a favorite teacher that that comes to mind?

SPEAKER_01 54:17

There are there are two teachers that um I think are really uh influential one Miss Mixon my seventh grade GT teacher who we had for basically half the day and who just encouraged us to be like uniquely ourselves and GT uh gifted and talented. She spoke at my wedding uh uh 30 years later and uh she's just an incredible inspiring human who like lives with so much passion for life. Um and then I also had my high school German teacher Frau Sessler very very important. And I think if I had not been um an exchange student in German for German uh to going to Germany as a high schooler I don't think I would be here uh because I would say I don't know like we can but really really briefly Alex um I tell the story and and you know I it took me a long time to realize that the reason I'm an educator and an educator of educators is because I came very close to taking my own life when I was 15 and didn't really feel like there was a space for me in this world as a queer Asian kid growing up without seeing anyone who kind of looked like me in in my part of Texas. And if not for educators who saw me and who believed in me and then who honest who literally took me around the world to see other possibilities, I I don't think I would I would be here. And so it is um I think a real honor of a lifetime to then continue to work with educators to kind of think about what is the type of learning that would allow young people to feel like they have a place in this world and that they can kind of shape uh the future they want to be in that that is like the root of the root and why I work in education.

Alex Kotran 56:01

Yeah it's yeah it's we I I had I I had a really rough time I mean growing up gay it was tough. I was in Ohio so maybe not probably a different ballgame than rural Texas but wasn't rural I was in Houston so in Houston yeah different ballgame than than Texas um yeah and I was bullied being a redhead is part I mean like I wasn't out but I was just like I was a redhead I was you know my parents were foreign I'd ever been and I was also just like a goofball so like I you know I was I attracted a lot of attention and I I there was there's a lot I mean it's funny if you would ask me the question there's I had my sort of go-to answer which is my history teacher um it's really the teachers that just that loved what they did right and I like I like my chemistry teacher Mr. Young was just loved chemistry and it made me want to love chemistry even though I was not a STEM kid. And so there was there's power in that but yeah if I go earlier there's so many teachers where there's just stuff that my parents couldn't have been helped help me with like one teacher literally watched me get picked on and like had me leave the room and then just like reamed the entire class was just like this is I don't even know what she said. But after afterward like that was literally the end of bullying happened after this teacher had like literally just like what are you kids doing? I mean I don't know what she said but afterwards people would come up to me and be like I'm so sorry and and I just like that it's not that my parents couldn't didn't want to help but that's just they didn't have the the visibility or the place or the agency to do something like that. And my mom is a my mom is a she's retiring this year but she was she's an assistant principal who handles discipline disciplinary issues. And so it's like literally every day she is like simultaneously these teachers are playing this role of like therapist and being in the receiving end of a lot of like the challenges that students are dealing with but then also taking action um and and it just drives me crazy that we're we we we don't even acknowledge like like the conversations are still so often how do we do how do we deal with these teachers who are they're not doing a good job in like the um and I thought post COVID that would change it didn't change I don't think um I think it's gotten worse. You think it's gotten worse.

SPEAKER_01 58:22

Yeah but none of what we just described could be replaceable by technology or AI at least not for a long time. And Rish and all of it obviously illustrates how deeply human learning is and how schools are and how important real authentic relationships with educators are and how life saving that can be.

Alex Kotran 58:45

And that's part of this whole idea of how do you how do we get out of teachers' ways in part teachers want to have I mean nobody gets into teaching because they want to use ed tech. I'm pretty sure I could be wrong but I mean most people are not inspired by the technology dimension of of teaching you get into teaching because you want to actually like help kids and like steer them.

SPEAKER_01 59:08

I think if as long as the ed tech it I think there's a difference between ed tech that might replace like a learning experience but I I do think that some of the ed tech that where kids are actually creators and using ed tech to kind of shape and design I think and they're actually agents of learning I could I think that could be really quite different. Because then it becomes like a an application of learning. I I do think to your point part of why there is this deep AI skepticism among educators is because there's been so much ed tech that's been sold to our schools tens of billions of dollars of ed tech that we can't actually point to it resulting in better learning for students and at the same time seeing all this investment in technology and an underinvestment in humans I think that is also kind of lifted up like what are what are we doing? And so I do think there's a rebalancing that's pop that's possible um and an equip and like a different type of language a different type of vision for what educators can shape right now.

Alex Kotran 1:00:07

Yeah especially when um there's the uh the this engineer that Meta hired for their superintelligence lab and I think that his compensation package was $250 million. Maybe it was one somebody got offered a billion and declined it um and and you look at the numbers for you know even national numbers I mean yes they're in the billions but they're only in like the billions not the hundreds of billions and I mean you see like AI companies raising hundreds of billions. It's like there there is capital out there and I think one thing you're illustrating though is you're gonna have better outcomes. You're gonna have more demand for the technology if we're able to invest

Elevating Teachers With New Roles

Alex Kotran 1:01:00

in the people and their ability to absorb it and channel it.

SPEAKER_01 1:01:03

Yeah I'm sure you read Sal Khan's article that came out last week where even for the founder of Khan Academy there's an acknowledgement that it isn't about just the tools and the tech it's actually the human systems that enable all of this to run and be and learning to happen. I also think about um this hundreds of billions of dollars we've been saying spending on AI data centers, as you know, and the idea that most of the graphical processing units, the chips that are going these data centers are going to be out of date in like two years, three years and we're gonna have to restock all of the chips. And so we're spending all this capital, but the returns on the capital have such a short shelf life and if you instead imagine that you invested those hundreds of billions of dollars in humans and instead like humans don't have the same shelf life obviously and obviously over generations learning kind of compounds and we know that just kind of having one like um career trajectory education attainment and family can shift generations of outcomes. And so I think there's just a very twisted view of what what's a good investment right now. And I know that I know we're supposed to be talking about AI, but there is like a little bit of rebalancing and actually investing in people that feels really needed right now.

Alex Kotran 1:02:17

I just I just wonder if even if you're like the a really big proponent of AI, I still think investing in people is it it it is in service of your goals ultimately because part of the challenge with all this AI investment is like well who's going to use the tools and you have a lot of these AI companies are the forward deployed have you heard of FDEs? No. This is the new term of art in Silicon Valley the forward deployed engineer. So you're anthropic Google OpenAI they have these engineers that will literally go embed into customers organizations and help them figure out how to like you know how do we how do you use pod or or Gemini more effectively? Interesting um that's kind of what we do in schools. It is and I'm and I'm kind of curious about because you meant you mentioned that you're doing some experimentation with the tool and you gave us like one preview um I know they're coming up on time but I mean maybe this is like with a bit of an eye towards towards the future and and putting the spotlight on leading educators. Yeah is there like do you see a role for this kind of like implementation sort of uh strategist that kind of plugs in like deeply understands the context but goes maybe beyond just coaching into actually like getting hands on like what is it like what are the tools that you should use and like how should you use them?

SPEAKER_01 1:03:30

Yeah so this is coming up a lot and I think the way we talk about it is almost a coherence layer around schools or a design layer around schools is what's true right now is if I'm a principal or or teacher leader, there are so many ed tech tools coming out. There are also new teacher leader or strategic staffing roles, different ways to use classroom educators, community educators. And then there's new policies, new mandates, new curricula and so there almost needs to be a layer that wraps around a school that helps them think through how do we what's the best of what's out there and how do we apply it to our context and how do we do that in a way that's really consistent and coherent with the outcomes that we want for our kids and the ultimate visions. And so we're starting to do more work around in our in our future space around model building and partnering with schools to think about well, if you were going to reimagine the literacy block or the math block in this new era, or even kind of think about what a more career connected version of learning might look like in this new era, like that's the type of new design project that we're taking on that's a little bit beyond traditional PD. It's actually more towards the school model building space, which, you know, obviously we had a school model building movement like 20 years ago, but the technology and the ideas are totally different right now. And so it seems like a a a necessary uh partnership opportunity again.

Alex Kotran 1:04:53

Yeah it's like why like I get this question a lot which is why do we we've been trying to transform schools for for decades. And there's actually a lot of philanthropy leaving the space because they're throwing their arms up in the air and they're like we're just not seeing the return on investment. So what's different about this moment and I think that is where I'm willing to give AI a little you know sort of a nod. It AI makes possible things that would have previously been really really hard. It's like it takes teachers sometimes years to sort of dial in their curriculum and they're coming to us and saying well are you gonna you really expect me to completely rebuild all my curriculum from the bottom up so that it incorporates PBL and like I don't have the I don't have the time for that. And it's like well actually that it is within the realm of possibility that AI could help not entirely I I think the the can it entirely fast track it no I think that's the other side of it. It's like how long does it actually take it saves some time. Yeah. But you really have to be embedded to understand what that looks like.

SPEAKER_01 1:05:52

W And I would offer like one of the reasons why I think we're able to do this work at leading educators is we have now a track record of moving Moving the needle for kids over 10 years. I and on just to be a little braggy for a second, we're the first professional learning organization in 10 years to achieve tier one what works clearinghouse data. So successful randomized control trials, successful causal studies. We now have, I think, about 25 causal studies on average, where schools we're working with are seeing nine months additional growth each year with students. And a lot of that is actually before the technology. It is actually the hard work of adopting better curriculum, aligning conditions, aligning coaching time, providing teachers with a lot of in-depth side-by-side support and continuing to study and learn and then iterate together. So on and do a lot of continuous improvement work with schools. I think we're going to continue to do that. And the idea is then okay, let's not just look at a curriculum and other conditions that folks are using, but let's expand it to the tech infrastructure and the AI that folks are using. And then let's be really rigorous and like let's learn alongside schools that are piloting this new work and help them see like what's working and what's not. And like let's be real about the barriers and also really celebrate the opportunities in an honest way. Um, that's that's what I think we're seeking to do.

Alex Kotran 1:07:07

And I can see that I can see why you're not necessarily running into the pushback that other like tech providers are, because I think at the end of the day, parents they want their kids to learn, right? I mean, I think and I think if you frame it as like, don't you want your schools to be like forward thinking about how to evolve school to meet the needs of the future? I mean, I think that is sort of an unambiguously, no matter how afraid of the technology you are or how excited you are. Um, I think the answer to that is always the same. I think it's it's but it's important for for folks to hear that like we are not here to rip it all out. And at the same time, there's like most of the stuff that's happening, or a lot of the stuff that's happening, you wouldn't necessarily even say is valuable. Because I think parents were also not happy about school before AI. It's not that everything was great and then AI came around and it's ruining it. It's like the status quo actually wasn't working um for many, if not most. So to close, I you're talking directly to a teacher or district leader, they want to start down this journey, they haven't partnered yet with leading educators. Um yeah, what's your piece of advice to just like how to how to just get started? Like where would you,

The AI Impact Cycle To Get Started

Alex Kotran 1:08:15

what should they do over the next couple of weeks or over the summer?

SPEAKER_01 1:08:18

Oh my gosh. Um, well, what we tend to do with schools is to figure out what is a student-facing challenge or opportunity and to start from that actual instructional challenge and to think about like, well, what actually can AI-enabled tools do to help with that particular challenge? And how might you actually gather evidence about how teaching changes and how student learning changes? And we call it our AI impact cycle. So we basically work with school teams over two-week cycles to think about, well, what's the barrier? How can we deploy some responsible tools to help with that barrier? And then how do we know if it's actually making teaching and then student learning better? And then like do that over and over again. And that's actually where we see the most improvement. And I'll give you a really simple example of that. Um, we're working with some schools in New York as part of New York City Reads, and there's pretty consistent challenges there where students are having trouble accessing the grade level text. It's great that schools have adopted a more rigorous curriculum. In some cases, students are then missing some of the background knowledge or knowledge that makes the text really accessible. Well, it turns out in the curriculum that have been adopted, many of the schools, the curriculum company has actually trained a large language model on the exact curriculum. And so you could now work with the AI to actually then figure out exactly how to scaffold the background knowledge efficiently for students and officially for teachers that gives more students access to that grade level reading. And so that's a very like specific. Here's a challenge, here's the like aligned technology. And then let's let's really test. Is it actually going to lead to deeper student understanding? And let's find ways to be honest about what's working and what's not. And I think that's that type of experimentation and commitment to really learning, uh, both ourselves and for students, that I think is at the heart of the work. I guess the last thing I would just say is I feel like I talk to folks uh who feel behind on AI all the time. They're like, oh my God, there's all this stuff coming out. And then you have to remind people like, hey, that didn't actually exist until January. So you're at most, you're like four months behind. This is not, none of this is like too late to start. And I think setting aside time for dedicated experimentation with folks who might just be a little more expert than you and to like feel safe to kind of play and understand. I think that turns out to be really, really important right now because we have to keep the technology is going to continue to evolve. And it's actually the ways in which we continue to learn together responsibly. That's what's going to be more evergreen.

Lifelong Learning Takeaways And Farewell

Alex Kotran 1:10:41

Lifelong learning. I mean, so we we are we are starting to realize that lifelong learning is not just a trope that we bandy about. It is it is like fundamental to this this moment, is like just keep learning, stay curious, feed your cu curiosity. But do in a way that's structured. I like this, I like the call out of like if you're curious about something, like actually rigorously explore that and try to understand it and then share that with your peers.

SPEAKER_01 1:11:08

And yeah, like maybe maybe on this podcast.

Alex Kotran 1:11:12

Maybe on this podcast. Alex, yeah. Um, thank you so much. Yeah, such a pleasure. Thanks for having me, Alex.