De-Mystifying AI with Eric Hudson

In which Dan chats with Eric Hudson, an educator and facilitator who's spent a lot time thinking and talking about AI in schools. Dan and Eric talk about why AI is so polarizing, the student-teacher divide on this new technology, and how schools can preserve critical thinking while not running in fear of AI. 

As always, I welcome comments and questions on Instagram @_dankearney_

Mentioned in the show:

School Leaders Innovation Forum: Los Angeles presented by Toddle

Learning on Purpose a Substack by Eric Hudson

Nine AI Messages for Students by Eric Hudson

The Augmented Educator a Substack by Michael Wagner

Verified: How to Think Straight, Get Duped Less, and Make Better Decisions about What to Believe Online by Mike Caulfield and Sam Wineburg

What can you do as an educator? A post and graphic by Vera Cubero

Music by Pierce Murphy

Transcript

SPEAKER_00 0:00

Hey listener, before we get started, I want to tell you about a free PD opportunity coming to Los Angeles. Toddle is hosting the School Leaders Innovation Forum at Pilgrim School on March 6th. This forum will bring together more than 40 regional leaders for a powerful day of professional learning. This in-person event will create space to pause, gain perspective, and explore how to lead with clarity in uncertain times. Yours truly will be on a panel exploring how schools can remain mission-driven and adaptive in these uncertain times, and there will be lots of discussion around lessons learned from AI and education. It's a great chance to meet some big thinkers in Southern California on March 6th. Registration link is in the show notes.

SPEAKER_01 1:05

That culture needs to change. And so a very simple move is to just create spaces where people can talk without fear of retaliation or punishment about their use of AI. Because a lot of schools I work with are in a space where any use of AI is a conversation about punishment or ethics and not a conversation about curiosity or learning.

SPEAKER_00 1:58

Case in point, I saw this story today about Virginia Tech and some other schools using AI in their admissions applications evaluations, which struck me as a tad hypocritical, and clearly struck anyone who read the article as something. A colleague was ringing me up soon after I reposted it to talk about what it all means. I don't know what it all means, but I do know that AI is here, and the more conversations we can have, the better. The more we can understand the better. The more we can examine how education is and will be impacted by this exploding technology, the better. AI is making us question a lot of our fundamentals. Let's take writing. Stephen Mintz over at University of Texas at Austin recently posted about a complete sea change in how he assigns writing in class, not take home, due to AI's ability to create an entire essay in seconds. No surprise there, professors and teachers around the country have been asking this question for a couple of years. But hell, John Warner, who's been on this show, wrote a book 10 years ago about why the five-paragraph essay is dead. So in some ways I feel like AI is simply spotlighting weaknesses in the system. Weaknesses some people already saw. But I digress. At this point, the only thing I can be certain of is that AI, even if it never gets any better, is here to stay. It's a thing that we all need to grapple with. Not necessarily love or even use, but we need to wrestle with the implications of the technology. And honestly, won't that grappling make us better at what we do? To help me grapple, I reached out to today's guest.

SPEAKER_01 3:38

I'm Eric Hudson. I'm uh I'm an independent consultant. I work with schools um on a variety of things on the teaching and learning side, um, especially as those things relate to technology. So most of my work right now is AI. Um and uh my story is uh I'm a teacher by training. I spent 12 years in the classroom, mostly middle school and high school English uh at independent schools in New England, where I still live. Um and then I spent 10 years at a nonprofit called Global Online Academy. I did a variety of things there. It's a nonprofit that does passion-based online courses for high school students, PD for teachers and school leaders. I did instructional coaching, ended up running the program there. And I left GOA about two years ago to go off on my own as a consultant.

SPEAKER_00 4:25

Eric makes so many great points in this conversation, including my favorite, that the onset of AI has a lot in common with the pandemic's impact on education. I hope you enjoyed it. Um, but I I recall back when I did uh we were following you, and you you were having just lots of great stuff uh on technology and pedagogy. And then I came back around to your things on Substack recently. You published an article called Nine AI Messages for Students. And I was so struck by how practical, useful, but also empathetic it was to this for students, to the students and helpful for teachers. Um and I'm wondering what what prompted you to write this?

SPEAKER_01 5:29

Yeah, uh, thank you for asking that. Um, I most of my work as a consultant is I go to schools and um I do a lot of professional development and consultation. And pretty soon after I started working on AI, I started asking if I could get time with students while I visit these schools. And some schools took me up on it, some schools didn't. Um and then I realized that teachers were talking to teachers a lot about AI, and students were talking to students a lot about AI, but those two groups weren't really talking to each other a lot about AI. And that really has to do with, you know, the impact AI is having. Like when ChatGPT came out, cheating machine, ban it, punishment, it's bad for you, like all that stuff. And so I think the stigma around it has sort of prevented communication. And so I wrote that post for two main reasons. Mainly I wanted teachers to understand the kinds of experiences students were having, but I also wanted students to try to see AI through the lens of what their teachers really care about, which is teaching and learning, authenticity, relationships. And so that article is just kind of like part of my ongoing mission to like have these two groups communicate to each other with each other more frequently and in about things that are not just cheating.

SPEAKER_00 6:55

Yeah, that's great. I actually had the privilege today. I was in a classroom, um, you know, I was kind of walking by and looked in the window and saw that one of our uh our eighth grade history teacher was talking about AI. I was like, oh, duck in here and see. And it was great. He was sort of, I think, trying to do some of the things that you were just were talking about, bridge that distance between him and the students, but also help the students understand how he sees AI and that he's not afraid of it. But that he thinks there's, you know, there are ways that they could smartly use AI as long as they're thoughtful about it. And and you brought up Chat and the cheating machine and everything that came out. And it you actually open your article by writing that uh we have this absolutism when it comes to AI that uh we tend to see AI in pretty stark terms. There are those that fear it, hate it, and those that are really embracing it. And you say that using AI doesn't make someone good or bad. Why do you think we've arrived at this such absolutism with AI? And how do you talk to students about this dichotomy?

SPEAKER_01 8:04

Well, really lightweight question. Yeah, yeah, just rolling and coming in hot. So um I think one of the reasons why there's a lot of sort of um absolutism around AI is because uh the strength of the emotions it provokes in people um is very powerful. Uh and so, for example, those who are really excited about AI get really excited about potential applications in the classroom, but also get really excited about, you know, things that have happened with the internet, like the democratization of assistance, you know, now everybody gets a tutor. Or the idea that um we now have this assistant that can create time for us to do things we care about, right? Whereas people who really resist it are like, you know, this um, you know, fruit of the poison tree, like it's a just a deeply ethically flawed technology, um, you know, things like environmental impact, algorithmic bias, like we shouldn't be engaging. And I think the strength of the emotions has kind of led people to um view how someone feels about AI as sort of an indicator of what kind of person they are. And I think that that is like not great for school culture, in my experience. Um, that it is deeply disruptive and deeply powerful and and complex and problematic in a lot of ways, but we have to find ways to talk about it that hopefully move us away from the way you feel about AI reflects your ethical belief system, right? Um, and that if your feelings about AI are in conflict with my feelings about AI, then we can't really have a good conversation about it. Like I don't think any of that is true, but that dynamic shows up for me a lot in my interactions, more at the adult level, frankly, than at the student level. Um, but I think it's just something to be aware of.

SPEAKER_00 10:13

Yeah, I'm curious because you spent so much time in schools talking with students and adults. What is sort of the difference or similarities in sort of how those two groups because uh part of it too is social media gives you polarized views, right? I I can scroll and see, you know, Jon Stewart screaming about the the profiteering of AI, and then two swipes later, it's someone saying, like, this is gonna transform the way you teach yourself something. But when you're out there in schools, what are you kind of hearing from students and adults? Yeah, what what the Venn diagram of emotions, what's that look like?

SPEAKER_01 10:56

So that's a great question. Um, I guess I would start with like what we know from research and data, which is students, in other words, kids 18 and under, are heavy frequent users, right? So like multiple studies will show that 80 to 85% of surveyed students have used AI for schoolwork, right? Um and anecdotally, when I talk to students, they place the number at like 90%, right? Whereas data shows that educators, really only around half of educators in surveys have said they've used AI at all. And so there's a gap in use, not literacy or fluency, but just literally in terms of like hands-on use of the tool. And so students, for students, it's just an embedded part of their reality in a way that, and I'm generalizing, in a way that it's not for adults. For students who spend more time online, they spend more time on social media in general, they have normalized it in a way that allows them to talk about it, that in a way that feels a little more grounded, frankly, than um adults. Students still ask me questions about are the robots gonna take over? Is this gonna burn the planet to a crisp? But with adults, the emotions that come along with that, I'm sorry, with students, the emotions that come along with that are kind of like worry, curiosity, right? Like a balance. With adults, it's um there's a lot of fear and anger, in my experience, right? Um, understandably so. I think when your profession is education and a technology like this comes out, there's a lot of things about it that would lead a person to believe that they're going to be replaced by artificial intelligence. And so I understand those emotions, but I think students just they haven't kind of like blown it up in the same way that adults do. And I and I truly think that's just because the kids experience it on a day-to-day basis more frequently than adults do. Sure.

SPEAKER_00 12:56

I think probably for a lot of adults, it just hasn't been demystified either. That there's just a certain usage uh experience that that I think would would help. You write about the difference between direct and collaborative use. And I think maybe a lot of adults' teachers fear AI because what they're picturing is direct use. They're picturing like the AI just is gonna give you the answer. I need this thing now, bam, I've got it. But you argue there's a pretty interesting distinction between direct and collaborative, and helping students understand that is really key. So maybe you could walk us through direct collaborative use of AI.

SPEAKER_01 13:36

Sure. Um, so those terms come from a study that the company Anthropic did. Anthropic is the company that owns the LLM Claude, and they looked at uh more than 500,000 interactions that college students had with Claude. And they did this whole long sort of study and analysis. But that piece of it really popped out to me, which is they categorize student use into two main categories, direct and collaborative. And direct is using Claude to get to a product as quickly as possible and in as few prompts as possible. So write this paper for me, make this slide deck for me, like make this annotated bibliography for me, right? All the things teachers fear. All the things teachers fear. And yeah, I mean, we'll get into this, right? Um collaborative was more process-oriented. So the prompts would look something like explain this concept to me like I'm in kindergarten, or my professor gave me this project. Can you break this down into steps so I can get a better sense of what I need to do? Or here's a draft of my work, can you give me feedback? And um what Enthropic found was that across college majors, about 50% of the uses were direct and about 50% were collaborative. So students were using it both to just generate products, but also to help them through processes. And the reason why I find that so compelling is because when you look at the emerging research on the impact that AI has on cognition, it is direct, it's frequent direct use that has a negative impact on critical thinking scores on assessments in those studies. And yet, in some of those exact same studies, there's like a subset of people who were taught how to use it more collaboratively and they didn't see any negative impact on critical thinking. And in fact, in some cases, critical thinking went up. And like that really aligns with what we know about cognitive offloading as a positive and negative thing. If you're using cognitive offloading to offload, that's bad. If you're using cognitive offloading to lighten the cognitive load so that you're more capable of completing a process, that's good, right? And the problem with us human beings is that evolutionarily we're lazy. And so our default is direct use, right? Collaborative use has to be taught and rewarded and encouraged, right? And I think that's to me, that's like my great anxiety about AI is like the riskiest choice is to do nothing, right? And so I think that if we if we want to do something, I think it's illustrating and modeling for students what collaborative use looks like and why it can be good and helpful.

SPEAKER_00 16:16

Yeah, I was speaking with uh Michael Wagner, who's a professor of design at Drexel University. And he has some very interesting thoughts about how um ironically AI could in some ways um elevate uh ancient ways of assessment, right? The oral exam, right? Because because teachers need to know what what students know. But he also made some really great points about you know, he he he he said that ignoring AI is not gonna help young learners through school and into the into the job market. But over-relying on it, as you say, also is no good. And it's the it's the person that can do both, that has the critical thinking skills that can collaborate. And he talks a lot about the design process and the collaborative nature and how AI can um assist in that. But in that conversation, I I remember thinking, this is amazing, but you know how to do this. You can do this with your students, and what you just said, Eric, about modeling. Teachers aren't there, a lot of teachers just aren't there. And I'm curious, like, how can we move teachers from the fear of my students who are just writing papers with AI to actually there's a whole other side of this that you can do with your students? You know, it that that this almost like a mindset shift.

SPEAKER_01 17:44

Yeah, I mean, I think what's the professional development look like? Like I for me, so when I'm a question. Yeah. When I run a workshop, I try to do, like, let's say I had three hours with a faculty, right? Like, I try to do three things. I try to present sort of, I try to paint a research-based picture of the impact AI is having on the world beyond school, because I think teachers need to be informed about the bigger picture issues. I have them discuss and debate use cases of AI. So I give them real examples of student use cases and teacher use cases that I've seen at schools and have them talk about the nuances of that use. So if a student uses AI for feedback on a piece of writing that they did themselves, is that good, bad, ugly, allowed, not like that's the sort of conversation. And then the third thing I insist upon is that we spend half those three hours using AI. Like, let's be in a chatbot. I'm gonna give you some tips on prompting, and I'm gonna give you time to try it, right? Because I think what you just said, you know, and what I'm sure Wagner's saying too is it's really only when you get in the flow with a chatbot that you start to realize number one, how much agency you have, but also number two, how much critical thinking is required to use it well. Um, that if you just have it spit something out for you, usually it's not that great. Usually that output requires your evaluation and feedback, right? And so it is it is an engaging experience to use a chatbot in the sense of you need to be tuned in in order to use it effectively. And that's the kind of thing that I really want teachers to feel is just the I want them to experience the way a lot of people who are regular users experience it, which is, yeah, sometimes it can make stuff for you, but most of the most effective users are actually pretty engaged in the back and forth with it.

SPEAKER_00 19:56

Yeah, how do those workshops go? I mean, I'm curious if you have any um examples uh of teachers who have, I don't know, maybe changed their minds or come in and and and maybe they're skeptical or maybe they're just novices and and then by the end of your your session, it's like, oh, okay, like they just see it in a whole new way.

SPEAKER_01 20:19

Yeah, I get definitely get that. Um, you know, I'm I'm still I'm still regularly surprised when I do a workshop about how few people in the room have ever even tried to use a chatbot. Um so that so I think even just that exposure is helpful. Um and yeah, I get a lot of positive feedback on, you know, we I mean, this is the whole thing with PD, right? If PD, if it doesn't feel relevant and practical, it doesn't happen. And so that's sort of what people like about the hands-on segment. But I also get very strong negative reaction. Like I get people who refuse to do it because they object to AI. Um, I get people who use it and they are surprised by how good it is, but their reaction to that is very negative. Like they're they go to the, oh, well, now I'm sure this thing is gonna replace me. A lot of teachers still live in the land of prompt engineering, where early on when ChatGPT came out, we were told it's not gonna perform well unless you craft a beautiful prompt. And like that's no longer true. And so they are very kind of surprised in a bad way by how good these bots have become at inferring intent, um, generating nuanced output. Um, and so I'm still seeing the full spectrum of reactions to the technology, which is understandable.

SPEAKER_00 21:56

There's something I think, especially those of us who are. Older. There's this the element with AI that it doesn't have any buy-in into who we are, you know, and it and it you you brought up this cheap signals idea from Mike Halfield that it that it has this fluency and this confidence and and even this care, like sycophancy, really, like that kind of presents this illusion, but it's not really there. And how how how can adults and students sort of interrogate their own interactions with the chatbot so that they're being careful to see through that and not get lulled in?

SPEAKER_01 22:39

You're like talking about the conversations I'm having with my husband and his relationship with AI. That's like like how are you interrogating because he he he's a he's a big chat GPT user and he really likes it. And I I actually his use cases of it are really clever and smart, but like I'm always like, are you like what's your what's your relationship with ChatGPT these days? Um but I think that book you mentioned Mike Caulfield, he co-wrote a book with Sam Weinberg called Verified, which is where that term cheap signals come from. It's an excellent book. It's about information literacy as a core competency. And I think that's what we should really be teaching students is information literacy, not just in the context of AI, but in the context of like this information, this content-saturated world we live in. Um but I also think we should be teaching them how to be um critical users. So when you use AI, um do you take 30 seconds to look at what it gives you and ask yourself, um, is this thing trying to please me? Have you ever tried to prompt AI and say explicitly to it, I want you to answer my prompt in a way that is critical? Um, have you ever tried to prompt AI to change its voice and persona to be more confrontational, to be less, because the default is sycophantic, fluent, professional. And so awareness of the default would allow you to prompt the bot to do something else because they will do something else. You just have to tell them. Um, I've trained Claude, for example, in my own personal use. I've trained it to be far more critical and far more confrontational with me because I use Claude mostly for planning and brainstorming purposes. And, you know, it's it's gotten really good. I mean, it swore at me unsolicited the other day, you know, um, as a way to as a way to like really introduce some friction into our interactions. And I and I think that like that that sort of literacy piece can be helpful to people who are trying to figure out how to use the tool most effectively.

SPEAKER_00 25:04

Thinking about safety, um, and there's just so many elements to safety and AI. Uh, you know, I think a lot of stuff in the sort of the ether right now um is the idea that AI is unregulated, it's essentially being controlled by just a handful of people. You know, what is really their interest in us, their interest is in profit. Um and when I when I hear those types of things, I that the analogy I usually think of is Instagram when in when it was discovered what Instagram knew about the effect it was having on teenagers and did nothing to stop it. And I and so I wonder what kind of where your thinking is around the ethics of AI, teenagers, schools, you know, and even let's make this question really big. But you know, schools using things like uh Magic School, which purportedly put some guardrails around it. Where do you landing on just the safety of young minds and AI?

SPEAKER_01 26:02

Yeah. Um, yeah, you really do come with the simplest questions to um so there is um Vera Kubero is um an educator researcher out of North Carolina. She helped the state um develop their guidelines, and she has something I can send it to you after um that I find super helpful on this topic, which is she's like, let's all acknowledge that AI is an overwhelming topic. Like it is so complex, it is so um in infiltrating almost every aspect of our lives. Let's acknowledge that. And then let's focus as educators on what's in our locus of control. And she has this great diagram of like, there's our locus of control, there's our locus of influence, there's our locus of concern, right? In our locus of control are things that we can act on and control right now. And things that are our locus of control around student safety are we can educate students about safe behavior online, we can provide them with tools like Magic School, like Flint, like School AI that have guardrails on them. We can build our adult literacy as teachers so that we are better equipped to have conversations with students. Then there's our locus of influence, which is maybe we as teachers can influence our school's policy or our district's policy on this, or maybe we can influence parents and how they talk to the kids at home about online safety. And then there's our sphere of concern, which is where your question came from, which is corporate consolidation, profit motive, like did like the lack of regulation by the government of any of these technologies or these technology companies to any effective degree. What's in your locus of concern deeply matters. It is very troubling. It's hopefully something that will be disruptive. It's pretty outside the locus of control of your average educator. And so I think when it comes to student safety, let's start by focusing on what's in our locus of control, which is educating students and providing them with tools that are safe for the students to use. If we're going to provide them with tools, um, some schools don't, which is fine. But I think if you do want to bring AI into your school, like then don't just give your kids permission and then think you're done. Like, really try to bring in tools that um are safe to use.

SPEAKER_00 28:25

Great. Um yeah, I look forward to seeing that uh that diagram. It sounds like it really kind of took my sort of messy question and was like, here it is. That's great.

SPEAKER_01 28:35

Not my first rodeo.

SPEAKER_00 28:40

But uh but on the you know, this this word safe, you also make the great point that chat and other chat bots are they can be really great uh essentially like safe spaces for trying out ideas and getting feedback, which really is interesting because for a lot of young people growing minds, that can be daunting. Sort of bouncing an idea off someone. Uh just you know what we might do professionally because we've had a lot of reps and we've made a lot of mistakes, and maybe we weren't proficient until we were older, but ChatGPT could give young minds a chance to get to that earlier.

SPEAKER_01 29:18

Adults and students, and there's like it's it's a double-edged sword in a lot of ways, because what people like about interacting with a chatbot on sensitive things is that it feels very private. It feels very psychologically safe. So, for example, you just started in an admin role, right? And so maybe this year, maybe you already have that this experience where like you have to have a really hard conversation with somebody you work with about something they've done. You have to provide them with feedback, you have to coach them, whatever it is. I meet a lot of school leaders who now use chatbots to practice for those conversations. So they describe the situation, they look for advice about what to say, um, they they pose what they want to say and ask for feedback. And that's not because they don't want to have the human-to-human conversation, it's because they're feeling this vulnerability about doing the thing right, right? And when you look at why students or people in general might go to AI for companionship purposes, it's because of that privacy, that psychological safety. Like, I just need a little comfort. I don't want to do a whole thing of like telling someone I'm having a problem and making it a big drama. Like, I just want to process something privately. And I think there's pros to that, but I'm sure you can imagine, like slippery slope, this is where we get into a lot of problems with relationships with AI, right? And so I think there's this like very tricky balance to be had around this question of privacy and what AI, this, the it's it's a new kind of space that AI has created around interactions and conversation and relationships, which is we had our own thoughts and we had our relationships with people, and now we have our relationship with AI. And that's it's like it's a very it is the part of AI that worries me the most, um, way more than sort of academic integrity or um questions about schoolwork, you know. Sure.

SPEAKER_00 31:26

Yeah, I know someone who maybe like a year and a half ago, so before sort of these outrageous stories of people like, you know, marrying ChatGPT or whatever were coming out who use it as a therapist, and she sent me, she's like, You have to see this. You I think I don't care if you see like what I said to it, you have to see this chat log. And it was, you know, it was like this 30-page Google Doc. And I was just floored by the quality of the advice that it was giving. And so, oh that's when I sort of was like, okay, this is way bigger than just like me at school and what kids are doing. It's it's really interesting. What you're in a lot of schools talking to a lot of administrators. What is something that schools, what are you seeing that schools are doing right about AI? And what's the biggest mistake you see schools making?

SPEAKER_01 32:25

So something that schools are doing right about AI is um trying to become a little bit more aware about the culture that surrounds AI and changing it. And so um you're not gonna get productive use of AI if the culture of AI your school is built on secretiveness and punishment and fear and suspicion. That culture needs to change. And so a very simple move is to just create spaces where people can talk without fear of retaliation or punishment about their use of AI. Because a lot of schools I work with are in a space where any use of AI is a conversation about punishment or ethics and not a conversation about curiosity or learning or whatever it is. And so I think schools, schools that are doing the right thing right now are like, oh yeah, we have to change how we talk about it before we're gonna be able to do anything productive with it. And I think that that's a helpful first step. I think a mistake schools make is a mistake that schools make with technology in general, which is they really see AI as a technology issue and not a design issue. So they purchase a tool like Magic School, for example, and they're like, great, done. Solved AI. We now have a FERPA compliant tool that we've given to our teachers, and they're gonna, they're gonna figure out how to use it, right? Whereas I think more productive conversations around AI are more about, you know, what about our assessments is no longer valid or productive now that we have this technology in the world? And what kind of design work do we need to do in our assessments? Like, I really do genuinely see AI as a pedagogy and assessment question more than it is like a technology question when it comes to classroom applications, right? Because the tool, I mean, you know this. Like when Web 2.0 was a thing in like the 2010s, everyone was like, oh, we're gonna like buy these 12 different tools that do this thing and it's gonna change how school works. But it was all driven by if the if the educator doesn't have an understanding of the pedagogy behind the tool, then it's it's not gonna do the job, right? And I think that's that's a mistake I would like schools to avoid is making a purchase rather than addressing the pedagogy.

SPEAKER_00 34:59

I totally agree on the pedagogy and assessment. I I thought I mean since ChatGPT emerged, I thought that teachers who were already sort of creating authentic assessments um suddenly found themselves kind of ahead in the game in a way, as opposed to teachers that were giving assignments that were just AI ready. And that's where I we were saying before the before we started recording that I'm in a middle school, and I feel like a middle school we're in a position where we can have sane conversations about AI because we're not losing our minds over the essays or the lab reports that are just being, well, we don't know. That's the thing. We're not sure if they're. And this kind of goes back to the conversation I had with Michael Wagner about having AI proof assessments, but that also can find a way to collaborate with AI. And I and are you seeing schools that are that are able to find that that middle ground? I guess what I guess what I'm wondering, Eric, is I guess what I'm wondering is is what I'm seeing and hearing sort of in my local environment about just the stress that high schools are feeling about AI. Are you seeing that too when you when you travel country?

SPEAKER_01 36:25

Totally. I mean, I think the the cognitive and emotional exhaustion of high school teachers is real. Like, you know, these are teachers, they these teachers did not get into education to be cops, right? And they feel like all they're doing right now is policing student use of AI. It's unsustainable. Um, not only is it exhausting for the teacher, but it's creating this like adversarial dynamic in classrooms that's really kind of demolishing relationships. You know, I hear so many stories from students about how they find out their teacher is running their stuff through AI detectors. So they start running their teacher's stuff through AI detectors because they want to know if the teacher's following their own rules, right? Like that sort of that stuff is just like so toxic for healthy classroom environments. Um, and you know, it's it's really hard because I think if we want to be specific about writing, I mean the essay is because of AI, the essay is no longer a reliable proxy for a student's learning. And so that gets at the heart of what happens in high school English classrooms, right? And so it's kind of undeniable that we need to do something else, right? But what that something else is remains unclear. And that's sort of the that's what's leading to, I think, a lot of the stress, a lot of the fear, a lot of the exhaustion is how do you how do you approach something where uncertainty is the name of the game and change is the name of the game, you know? Um, it's really hard.

SPEAKER_00 38:11

It is. Because it's not just disrupting the knowledge base, uh, you know, or or what's the work students have to put in to learn something. It's disrupting the rhythms of the school day, the rhythms of the class teachers used to doing this, then this, then the students do this at home. Whatever the flow they have, that flow's been totally broken by what's at their students' fingertips.

SPEAKER_01 38:36

Yeah, yeah. And it's it's it's similar, it's similar to because I was at Global Online Academy during COVID. And so working in an online learning organization during COVID, you really, you really saw, how I really saw how the impact that a sudden, uninvited disruption on teaching has on adults and students. And what's happening with AI really reminds me of that, which is this is not something that teachers asked for. It's um in total misalignment with, like you said, their their sort of expertise in what they know to be true and valuable at teaching. And they are being forced to adapt. And that is like the worst possible context for meaningful change is being forced to do it just because it's so exhausting, right? And I think that, you know, that that to me feels very real. I'm like full of compassion for that part of it. Yeah.

SPEAKER_00 39:39

COVID is such a good comparison, you know, because I remember those, that transition to remote learning. Everyone did in some form. Some schools did it really well, others not as much. And a lot of teachers, in the same way, there was the it was the unknown, the fear of the unknown was really strong. And I can remember being so proud of those teachers that were so worried or resistant to the idea of this shift. And then once they embraced it, they got so good at it. But it was demystifying it, it was it was trying it out that made the difference.

SPEAKER_01 40:13

Yes. And it was, it was working within the constraints that you've been given, right? Like I think it's not fighting the constraints, it's acknowledging them and then trying to figure out how you can be who you want to be within those constraints. And I think that, to me, that was my experience with the most successful teachers in COVID. And that's my experience with the most successful teachers with AI is, you know, I was on a call with a couple of teachers from a school in Canada that's doing some super innovative work with AI. I can talk about it if you want. But the this English teacher said something that has really stayed with me, which she said, um, if I could put AI back in the box, if I could push a button and eliminate AI from the world right now, I would do it. But I'm also not gonna sit here and pretend that my work is not being transformed by this technology. She's like, I'm gonna get ahead of it. I'm not gonna wait for someone to do something, right? And I think that commitment to your own agency is just like incredibly important right now.

SPEAKER_00 41:20

Yeah, that's it. You know, the constraints within COVID, it's leaning into those constraints. It's like, how could these constraints actually be an opportunity? We saw a lot of that during COVID. And I think that's a great way of putting it. That that where's my agency? And I'm gonna assert my agency. I mean, let's uh, you know, I'd I would love to hear about a success story. Either it's a student who did something really compelling or it's a school that did something really innovative. What what what would you tout as something you've seen that's been just really cool?

SPEAKER_01 41:51

I mean, I find what kids are doing, what kids are doing around, like when I I when I sit with students, I'm I say, if you had to describe the job that AI does for you, what what would you say? And the number one thing kids say that they're willing to share with me is um tutor, right? And they, because they, especially when if you go to an independent school, private tutoring is just a big thing for a lot of well-resourced families. And so a lot of students who don't have access to private tutors or don't have access to frequent and deep parental support have figured out how to use AI to play that role in their work. And they do really clever stuff that I would argue is far more collaborative than direct. Um, having AI quiz them and test them, having it explain unclear instructions in a clearer way, using it as a tool that allows them to navigate, especially high school, in a way that helps them kind of manage all the moving parts of being a 16-year-old at a high-achieving school. Like I'm really impressed by students who think of it in that way. Um, and I think with teachers, I'm very, very impressed by teachers who have decided to take an action research approach to AI in their work, which is I'm not gonna burn the house down, nor am I going to pretend it doesn't exist. I'm gonna pick one unit, one project, one thing I do, and I'm gonna beta test an application of AI. I'm gonna see if it actually does the thing that all the hype mongers online say it does. And they test it for themselves and they are very transparent with their students about what they're doing and why. You know, I give you a very concrete example, which is um, you know, a physics teacher I met who was teaching uh an astronomy course, and he taught the kid because and there's these enormous public data sets freely available online about exoplanets and moons. It's all funded by the government, so it's Freely available online. And he taught the kids how to use Chat GPT to code little apps that would allow them to analyze and visualize that data. So if the students were interested in like the relationship between mass and luminosity in exoplanets, like they could use ChatGPT to vibe code an app that would allow them to sort of generate graphics, right? That kind of thing is the stuff that gets me most excited of like a teacher is not wondering is AI going to ruin my class and/or is AI going to change my class? It's a teacher who's like, oh, I wonder if AI can take what I do with students one step further, right? Or is it that is it does is can AI do something that I've always kind of wondered about in my own teaching, but have never really figured out how to operationalize or concretize. Like that, that's where I get most excited about AI is teachers using it in service of their own goals and their own values.

SPEAKER_00 44:59

Yeah, you uh, where is this? Um you had a great line in here about, of course, now I can't find it because rather than write it down, I was like, I'll certainly find it. When I'll hear is when you said um, you wrote, you'd let you say to students, imagine how AI can help you do the things you care about, not avoid doing the things you don't. And I just think that's so such a great way of framing it that it's not about it being a shortcut to get just this thing done. What are the things that you enjoy that you're passionate about? And it can be a tool to help you learn more or do more.

SPEAKER_01 45:34

I mean, I learned that lesson from students, for sure. Like when you talk to students about their AI use, students are using AI for school. If they're using AI in inappropriate ways for school, they're just trying to create time to do things they care about, right? So they're using AI to write that English paper so that they can have more time to, like one student was like, Yeah, I use AI so that I can enter international coding competitions, right? Or another student who's super into like web design, who's using AI to create wireframes for different websites they want to build, or a student who's interested in graphics, like any, whatever it is. Students are using AI as much as a shortcut as they are to support things they care about. Um and I think that they they do have an understanding that certain use cases of AI are probably not great, you know, but but it it's very much it's very much what we know from research about why students cheat. Students cheat because of a perceived lack of time or a perceived lack of ability or a perceived lack of relevance, right? And so if AI can make time for them to do other things that they see as more relevant or more meaningful or whatever, like they're gonna do it. And then they're gonna use AI to help them do those things. Like it's so it's a both and for students in a really interesting way.

SPEAKER_00 47:03

Ties back to that big word agency that you use. And you close the piece by talking about Wally, the great film, and you know, this world in which humans have delegated their agency to essentially AI or something like it, to be comfortable, entertained, utterly helpless. And you you finish the piece by saying, How do you want to be a human in an AI world? Which is such a great question uh for for us all to consider. But for those who, as you say, to students, AI is just a thing in their life. It's not particularly novel. It's a particularly important question for them to grapple with.

SPEAKER_01 47:41

Yeah. Um in a couple of weeks, I'm doing uh a keynote at a school. It's a day-long sort of student symposium around AI, and I'm speaking to the uh group of a couple hundred eighth to twelfth graders, and that's the question they want me to answer in the keynote is what does it mean to be human in an AI world? And so I'll let you know what I come up with. Yeah.

SPEAKER_00 48:10

I think we all would love to know Eric was. Yeah, yeah. You can just let human writ large know what.

SPEAKER_01 48:17

I mean, I I do think, I mean, we'll keep this light because I think I do think this is that's one question where actually culture can be really helpful, like movies and books. Because when I talk to students about AI, I talk about Wally because they've all seen it. These kids have, this is a generation that grew up with Pixar, and they get it in that way. Like the the thing that is so poignant to me about that movie is that it's a love story and an adventure story, but the robots get the love and get the adventure because the humans have given it up. And I think that there's like, there's just something so kids, I think students get that. And when I talk to them, talk to older students about, for example, the movie Her. Have you seen that movie? Walking Phoenix falls in love with the voice of Scarlett Johansson. They get that. I think they get it. They uh in my experience, those movies that take AI down to a human scale, which is how do we think about relationships? How do we think about trust? How do we think about agency? Like, let's move away from like matrix terminator stuff and move towards um stories that talk about AI's relationship to the things that we care about as human beings and how easily we can give up those things just because this tool makes it easier. And like, how do we just choose the things that are hard? Um, and I think that's that's an important. For me, that's like the the lesson in all of this is like we're just gonna have to get more intentional about choosing the things that are hard.

SPEAKER_00 49:49

So wonderful. Eric, I want to thank you so much. The article, the post, nine AI messages for students. It's in the show notes from your learning on purpose substack. Uh really cool. I I love you're able to crystallize so many big ideas into uh into really uh practical terms. Uh really helped me a lot, actually, just kind of seeing it through a lot of these lenses that you uh laid out today. Thanks so much for joining me.

SPEAKER_01 50:18

Thanks so much for having me. This is really fun.

SPEAKER_00 50:22

Huge thanks to Eric Hudson for joining me on the show today. The more we can talk about AI in education, the better. I just I said that at the top of the show, and my conversation with Eric just reiterated that for me. You know, at my school, we're just introducing conversation at our staff meetings. As teachers are all different points using AI, understanding AI, being comfortable with it, and we're just having uh chats in our meetings about what it all means. It's not an endorsement of using it, but it's also not shying away from it. We're leaning into what AI uh is going to mean for us in the near and long term. And I encourage you to do the same. Uh, if you're on Substack, if you're on social media, you are seeing so much about AI, but I would caution you to remember that a lot of those are the strongest emotions. They are people on the extremes who are either very hostile to AI or who have really endorsed it often for monetary purposes. So at your own school, what's going on? Are people using it? Are people against it? Are they for it? I'd love to hear some comments. Drop them uh on Instagram, um, and uh let me know. Is your uh is your school leaning into AI? Is it avoiding it? Uh do you think we need to be using it? Do you think we need to be uh wary of it? All right, happy 2026, thanks for listening, and uh check it out next time.

SPEAKER_01 52:06

Shadow of uncertainty.