Do Chatbots Really Belong in Schools? w/ Tom Mullaney

Generative AI is making its way into many parts of society, and schools are no different. Tom Mullaney joins Paris Marx to discuss how generative AI has been adopted in K-12 education and the many concerns it presents for students and teachers.

Tom Mullaney is a high school social studies teacher in the suburbs of Philadelphia.

Tech Won’t Save Us offers a critical perspective on tech, its worldview, and wider society with the goal of inspiring people to demand better tech and a better world. Support the show on Patreon.

The podcast is made in partnership with The Nation. Production is by Kyla Hewson.

Also mentioned in this episode:

• Here is the New Yorker article on AI in schools.
• For those looking for a refresher on Weizenbaum and ELIZA.
• Here is the paper “On the Dangers of Stochastic Parrots: Can Language Models Be Too Big”.
• For those curious about the Canvas breach.
• Students have been booing pro-AI speeches and AI presence in graduation ceremonies.
• xAI is facing a lawsuit for polluting Black neighborhoods.

Transcript

SPEAKER_01 0:00

I don't remember being an elementary school student and thinking, like, okay, this is what I'm doing to learn how to do a job, right? You're there to be educated. And we as a society, as a K-12 profession, we need to have that conversation and figure that out. I personally don't think of education as job training, but I think so many people do now that it's a situation that we need to resolve. And I don't know that thinking of it as job training is the most child-friendly answer to that question.

SPEAKER_00 0:50

I'm your host, Paris Marks, and this week my guest is Tom Bellaney. Tom is a high school teacher who teaches social studies in the suburbs of Philadelphia. And Tom has been writing a lot about generative AI in schools, you know, has been talking about this. Certainly, we ran into one another on Blue Sky, which is where I started to get to know him and his work and get to know his opinions on this, right? And I was thinking about doing uh an episode on generative AI in education for for quite some time. And then I was reminded recently uh when there was a New Yorker article, but also some listeners asked for an episode on this. And I said, Yeah, I need to reach out to Tom. We need to finally kind of do this episode, understand what is going on, had this discussion. And I was really happy he could make the time because I think that we had a great and wide-ranging chat where certainly we talked about generative AI in education and what that means for students, how students are using it, but also what it means for teachers and school districts. And we extend that conversation as well to talk about the use of digital technology more broadly, how that has been pushed for the past couple of decades, you know, what that actually means for schools and teachers and students, um, you know, going from kind of Google and Google Classroom and the reliance on those sorts of tools, but also, you know, how teaching and how education has changed over the past number of years, um, and and how that's tough on both students and teachers alike. And these tech companies really take advantage of it to push their products and make it seem like they're going to make things easier and more efficient while really, you know, they're just getting data, getting information, getting money, certainly. Um, and, you know, does that really work for us? Um, you know, and and we end by talking about the growing backlash, right? Not just to generative AI in schools, but also to screens more broadly and what that is doing to education and you know, where these things might be going. I think it's fair to say we both see the see this as kind of a hopeful movement uh to try to uh bring education away from the technology and actually focus more on on critical thought and and you know the general principles of education. Um so yeah, uh I guess that's all to say that I had a really good time talking with Tom. Um, I learned a lot from him, and I thought it was really insightful. Like I think that you're gonna learn a lot as well. And, you know, if you're a parent, you're probably more familiar with this than I would be. But hopefully you think I kind of asked the right questions. And certainly even if I didn't, Tom has a bunch of examples and a bunch of insights to share with us. So I think I think you're going to really enjoy it. As always, if you enjoy this episode, you can leave a five-star review on your podcast platform of choice. You can share the show on social media or with any friends or colleagues who you think would learn from it. And of course, you know, just a reminder that we do have a video version now that is available on YouTube, Spotify, and I think we'll be on Apple soon. I don't know 100% how that works, but we'll we'll see when that starts or if it has started already. And of course, if you, you know, do want to support the work that goes into making Tech Won't Save Us every single week, so we can keep having these critical, in-depth conversations, exploring so many different parts of the tech industry. You can join supporters like Michelle from Montreal or Belinda in Baltimore, who also wanted me to say, long live Logan Hollinger. And you can do that by going to patreon.com slash tech won't save us, where you can become a supporter, get some stickers if you support at a certain level, and certainly get access to premium episodes, which I'll have more information on soon. So enjoy this week's conversation. I think you're really going to like it. Tom, welcome to Tech Won't Save Us. Paris, I'm a longtime fan. Pleasure to be here. Thanks so much. I obviously really appreciate that. And we've been talking for a little while and have been meaning to do an episode on AI in education. And, you know, I saw a New Yorker article recently and like it brought it back to me. And I also had some listeners kind of asking for an episode on this subject. And I said, yes, I was talking to Tom. It's time to finally do this. And, you know, I'm happy to have you here to chat. And so I wanted to start just by asking like, you know, generative AI in schools has become a big topic of conversation. I feel like even just, you know, the kind of digital technology use in schools is becoming something that is more debated broadly. But I wonder ChatGPT was announced in November of 2022. How quickly did you start seeing these chatbots and things making their way into schools?

SPEAKER_01 5:15

So they made their way into the discourse about what students should be doing almost immediately. And what I mean by that is by the spring of 2023, there were multiple books available that promoted generative AI in K-12 schools. Think about that. Spring of 2023, multiple. And the authors of them then go on to keynote at education conferences. So that was immediate. What was not immediate and what's still lacking is kind of that balance that, okay, well, what do we know about generative AI and how humans use it? So what did we learn from the Stochastic Paris paper? What did we learn from Eliza and Weissenbaum? That has been missing from the discourse in K-12 the entire time. So, yes, very, very immediate. And schools, the animating principle behind generative AI and K-12, both I think amongst that discourse early on about how great this is going to be, and as schools say, all right, we have to incorporate this because it's here and it's inevitable. It's fear of missing out. Nobody wants to be an archaic dinosaur. And so fear of missing out has been the animating principle behind a lot of this.

SPEAKER_00 6:30

That's so frustrating. And but it's wild to hear that there were books available so quickly. And unfortunately, it doesn't surprise me that then these people go on to kind of speak at the conferences and to promote it because we see this cycle again and again, right? But it did feel notable with generative AI that very quickly, if you listen to people like Sam Altman, the potential educational uh, you know, implementations of this technology were something that they were hyping up from the very beginning, right? You know, kind of this notion that the chatbots can eventually become teachers or teachers' assistants and, you know, can kind of improve access to education and all this kind of stuff was like part of the grand, uh, you know, amazing things that generative AI is going to do, which I'm sure kind of affects the discourse around, you know, education and and how teachers and school boards and things like that are thinking about it and talking about it.

SPEAKER_01 7:25

So a lot of K-12 professional development and discourse is focused on people like Sam Altman or even say Ethan Mollock, have you heard of him? He's that UPenn professor who's always promoting generative AI and always saying, this is amazing. This is he's always promising that it's going to revolutionize everything, and then you know, it'll just do it the next time something is released. And voices like that who have a bias and an interest in us thinking that generative AI is truly transformative, those voices have dominated the conversation and they're cited as experts. As a socialized teacher, I want my students critical thinking. And we need to have critical thinking as well. And so if we're going to listen to those voices, we need to acknowledge their biases and why they have an interest in us thinking a certain way. And we should also listen to voices such as Dr. Emily M. Bender, Dr. Tim Nagebru. Let's listen to all of that and let's evaluate and critically think about that, and then let's proceed with what we want to do with technology.

SPEAKER_00 8:25

It is kind of wild how even in institutions like you know, schools and universities, that that kind of critical approach to the technologies, whether it's generative AI or any others, doesn't seem to be at the forefront of the discussion early on, right? You would think that if any kind of institutions in society are going to take a hard look at the technologies being proposed for society and certainly for their institutions in particular, you know, what's it going to mean for schools? What's it gonna mean for teachers? What's it gonna mean for students? Uh, you would think that they would be on the forefront of considering those questions. And it's wild that that is often not the case, right?

SPEAKER_01 9:03

So I'll give you an example. So at Google's I.O. event, Google I.O., they start talking, they talked about something that's been basically debunked by now, but this idea of singularity or AGI, that generative AI, large language models that just generate text, that's all they do, they're going to achieve consciousness and humanity, right? And be smarter than humans. And so uh I've seen in since that at least two ed tech influencers talking about that and saying this is why this is so compelling. Wired magazine, Wired magazine is not super anti-AI, but they quoted that stuff and said that's the most unhinged stuff said at this conference. So it's kind of like, you know, we're hearing that computers are about to be human, but the tech media who have been somewhat supportive of generative AI are saying, come on, that's ridiculous. So that's the problem, is that I think in K-12, we're hearing this message that, you know, if I didn't know any better, I would say, oh my God, like if the computers are gonna be smarter than us, then yeah, we have to do something. Um, rather than saying, all right, this outputs synthetic text based on prompts that we enter. So what is what are the consequences? You know, like that's what we should be saying. But instead, it's no, we're gonna have superhuman computers, right?

SPEAKER_00 10:25

Unfortunately, we see the same problem in in government and and other kind of institutions and and you know, places of governance that that should be thinking through these questions, but instead they buy into these narratives coming from the tech companies, just as you're saying, right? Whether it's around singularity or or whatever else. And so I want to dive in and talk about generative AI, but I wonder before we start uh you know to really dig into that conversation, obviously there have been concerns around how these tech companies have been relating to schools for a while. And I wonder, in your view, how did the rollout of generative AI in schools build on some of the existing concerns that existed with, I guess, the influence of these companies on what happens in K-12 schools?

SPEAKER_01 11:08

Okay, so that started. So, what got me down the path? Because for years, I was someone who incorporated technology and promoted technology in schools. And I do think there are some real valid uses for it that make teacher lives easier, student student lives easier, that make things more inclusive. I think there's absolutely some things to be done there and said, But what happened was, you know how brands use influencers and things like that? That's really what, especially Google, they led the league in that in like the early to mid 2010s. They were excellent at that. And then they're offering their stuff to schools for free. And it wasn't until later in the pandemic when schools really were dependent on it that they started charging for it. Um, but that's really so ridiculous. Yeah, that's how it really happened. And I would say I would encourage schools, I've actually talked to some administrators and they say, look, we have to have Google, we have to have Microsoft. It's just for all of the things that happen student information systems. I would say, like, look at things like like ellipsis. There, that's an that's a word processing tool that's available for free. I would say like pilot that or have just a few like advanced teachers play with that and see if they could use that as opposed to say like a Google Docs or something like that. What these companies did Google, Adobe, they were very adept at conferring prestige on people for using their products. And schools can be a very isolating place. You know, I might do a lesson about the New Deal, and maybe it just home run just is amazing. And maybe my students know that, right? And maybe if a principal is there observing, they know that. And then besides that, not much outside recognition. So these tech companies were very smart about, and what happened was by the time we get to like pre-pandemic or pandemic, I would get emails from teachers or from tech leaders at schools, and on the signature, there'd be 50 badges of 50 ed tech companies I had never heard of. Here, we'll give you a badge. And my thing was I was one of the prestige of the companies that my friends outside of education would know. So I I, you know, as I look back, I promoted those companies, not these smaller ones, but I don't, I'm not super proud of that. But um we've all made mistakes. It's okay. Yeah, yeah, definitely. Um anyway. So they use prestige to get in and get teachers on board, typical influencing, uh brand ambassador stuff. And that I the people I see in those spaces, they are not, I think, taking to heart the lessons now that are coming out of families being very concerned about technology in schools. Like they're not taking that to heart, or at least considering that, hey, maybe we need to like really think deeply about this. You know, I've definitely reconsidered a lot of my beliefs over the last two, three, four years. And I think that's only for the best for sure.

SPEAKER_00 14:06

Yeah, and and I think what you're saying there is something that a lot of people have experienced over the past number of years, right? You know, kind of a lot of us bought into the ideas that the tech company sold us, whether that was about implementation in schools or or how these technologies should be used in our general lives. And I think there are a lot more people who are kind of asking the question as to whether that made sense and changing their practices and changing their views on those things. So I don't think you're like outside of the norm on that at all, right?

SPEAKER_01 14:33

Well, thank you for saying that. It's definitely something I've I've struggled with. And as as I've returned to the classroom in the last year or so, I'm kind of like, okay, well, maybe I can like do some good here, you know?

SPEAKER_00 14:45

Yeah, no, I I hope so. I I feel like hearing you talk about that as well, like it brings back a story to me that that always stood out when I think about these tech companies and and education. And that is how back in the 1980s, Apple was one of like the first movers to get Macintosh comput, you know, to get computers in general, but it really wanted to get Macintosh computers into schools, and in particular in California, where it was based. And like I think it just gave a ton to schools for free, if I'm if I'm remembering correctly, or they were very cheap, but I think they were free. And and really the goal was like to introduce people to the products, to get people depending on the products, to get people used to them, you know, potentially cultivating future customers, of course, but also just kind of building in that dependence early in a place where you know you have a bunch of you know employees who are who are going to be using these things, but also, you know, you're kind of getting to people when they're young and they're developing an association with the brand. Or, you know, if you're thinking of like Google or Google Classroom, like they're actually getting a Google account very early in their lives. And I'm sure the hope is that they're going to then carry that through, you know, much of the rest of their lives, right?

SPEAKER_01 15:53

So Google and Google Classroom, they, you know, it's about 2015 and I'm or actually sorry, 2014. And so when that happened, and I will say having an LMS does give you, you know, students are absent, bam, they just go to the LMS. Like there are there are definitely. Thank you. Sorry. Learning management system. And there are others. One is kind of notorious right now because it just had a ransomware attack, uh, Canvas learning management system. So what's happened is yes, everything's become Google Classroom. And the funny thing is, is that when the pandemic at the onset of the pandemic, Google Classroom became almost ubiquitous. But when in the Apple store, students would review bomb it. They would all give it like one star. And I don't even know. I think at a certain point it was cut off. But what does that tell you? Like, what does that tell you if students are doing that? Maybe we should like listen to that rather than consider that bad behavior? Maybe that's like honest testimony. Um, so yeah, it became basically ubiquitous with the pandemic. I remember thinking before the pandemic, you know, this is like a third, like maybe near a half, but not quite. And then at the pandemic, it was like, well, we're in Google Classroom, and there's, you know, there's Microsoft schools, and so they were they might use another learning man management system like Schoology or Canvas. I will say this, having worked with all of them, Google Classroom, Schoology, Canvas, they're not the most exciting things to ever work in. They're not amazing. Uh, even just if you use Slack, and I've used Slack for different things when I was consulting, Slack is so much more fun and like things you can do. But I think there's kind of an inherent mistrust of kids. So it's kind of like, okay, well, don't let them like do certain things, don't let them like emoji react in certain contexts. Uh, just little things like that, that it's kind of unfortunate. Um, that's how kids I a lot of ways, like I think of how I experienced technology and how I for years technology has been like, oh, this is like better. This is like better than what I could do before. Um, even like Adobe Photoshop in the 90s, when I first was exposed to that, I was like so excited. Like, wow, that's amazing. Um, and now it's a subscription thing, and and people who use it are resentful about that, right? Um, this is the way our kids are experiencing technology for the first time on 11-inch Chromebooks. So I wonder if they get to, you know, they get to my age, they're just like so over it. They're so like, yeah, technology is for unfun things. Maybe, maybe that'll happen. I don't know.

SPEAKER_00 18:29

Yeah, and maybe we maybe it's even uh in high school when they're when they're reaching that point, as we see these like uh Luddite clubs and things like that in in Brooklyn and and whatnot.

SPEAKER_01 18:38

Definitely see that. I mean, we know about the backlash about generative AI and the commencement addresses, but I think I see that among students too, where students are wanting to get analog. And of course, the I don't know if if Canada is experiencing this, but the big trend now, the big thing that kids, high school kids are doing in the United States is playing hacky sack. I mean, don't talk about the most untechnotechnological thing you can do, hacky sack. So I think there is something to that.

SPEAKER_00 19:06

Interesting. That's really interesting. I feel like when you're talking about the learning management systems, one of the first things that came to my mind is like, I think they came in just like sort of after I finished high school and stuff, like a number of years after. And I was like, and I but I think my brother had them, or or one of my brothers did. And I was like, if my parents would have been able to look up my grades that easily, I would have hated everything. Like I would have been so mad.

SPEAKER_01 19:33

So that's and that's um that's even a different system called a student information system. And sometimes learning management systems play nicely with student information systems, and sometimes they don't. And you're typically your student information system, which is this thing called, usually called power school, and that's a whole other thing. Parents and guardians have quick access to grades. So, yes, uh, I would say, I always say it would be my nightmare to wake up and be one of my students. Not that I think I'm that better teacher, but what we ask of students today in this digital age, the amount of work we give them, their life experience, what they're looking at with generative AI and their future and just the world we're giving them, it would be a nightmare to be in that position. I I feel nothing but solidarity with kids today, for sure.

SPEAKER_00 20:23

Yeah, I I think I think that makes a lot of sense, right? And I I want to start moving into talking about generative AI specifically, but but hearing you say that, it actually brings to mind something I've heard a lot about universities, right? That, you know, part of the reason that generative AI and that these tech companies can so easily kind of move into universities is that they're taking advantage of some of the flaws that have already developed in kind of the model of the university, right? What university education has become, you know, how it has been transformed by administration and consultants over so many years. And, you know, you don't have the same student experience, you don't have the same kind of expectation from a university as you might have in the past. You know, you're maybe trying to get a degree to get a job rather than develop your critical thinking skills and all those sorts of things. And I feel like, you know, in that sense, the companies can take advantage of this flaws that already exist in the model. And I wonder then in K-12 schools, if you think that how K-12 education has has kind of developed over time or has evolved uh because of financial pressures and all this kind of stuff, if the companies are are also taking advantage of those problems to kind of move in.

SPEAKER_01 21:31

All right. So this is a reckoning we have to have in K-12. Either K-12, and I'm talking kindergarten, you know, what is that? Five years old through twelve through 12, which is 18 years old, either it's job training or it's education. Because I hear so much about well, we need, you know, corporate America needs this, and kids need these job skills. And I'm sitting there thinking, does a fifth grader need to be ready to work it for a corporate? Operation? This is okay. What are we getting at? And the justification for so much of this technology is that students need real world skills and they need to be trained for employment. I don't remember being an elementary school student and thinking like, okay, this is what I'm doing to learn how to do a job, right? You're there to be educated. And we as a society, as a K-12 profession, we need to have that conversation and figure that out. It I personally don't think of education as job training, but I think so many people do now that it's a situation that we need to resolve. And I don't know that thinking of it as job training is the most child-friendly answer to that question.

SPEAKER_00 22:47

I I couldn't agree with you more. You know, I think it causes a lot of issues in K-12 education and also in university education, right? Um with how these systems have have transformed and how they've been so reoriented toward preparing people for the job market, employment, all those sorts of things, rather than which is not to say cultivating other skills is not done. Um, but it feels like it's not the focus in the way that it should be or that maybe it it was at a certain moment, you know?

SPEAKER_01 23:15

Right. And critical thinking amongst our students, especially because I don't know that being iconoclastic and being different and thinking, you know, outside the box to be cliche is really what employers want. You know, they they say it is, but is that really what your an employer wants? They want a bunch of employees who are gonna do things differently, go their own way, disagree. Is that what they really want? But I, you know, I thought education was for critical thinking. So there's a bit of a conflict there. And I hope that's resolved in, like I said, a child-friendly way.

SPEAKER_00 23:50

It makes me think of all those articles that are like Gen Z workers are complain so much on the job or are so willing to like leave their jobs if they don't like their boss or whatnot. And it's like, I feel like when you're talking about is it critical thought or is it job training? What do employers want? Like, that's the first thing that comes to mind, how they're annoyed if people do push back, right?

SPEAKER_01 24:10

Like self-advocacy in a lot of contexts is not valued. And I would hope that students learn self-advocacy. I'll can I give you an example? So the UFT recently promoted a generative AI app and they told a story about it. Now, this is the United Federation of Teachers, Teachers Union, one of the teachers' unions in the United States, and they tell an anecdote with a student who says, you know, I don't want to bother my teacher so I can ask this chat bot a question. That is the bleakest possible outcome. As a teacher, when a student asks me a content-related question, that is the highlight of my day. That is what I want. I want them curious. I want them to think of me as a support and a resource. The last thing I want is them saying, you know, Mr. Molini doesn't want to hear this right now. Let me ask the chat bot. That is not the outcome. And that the UFT of all people is promoting that. That is really, really scary. And again, like we really need to reconsider the role of technology in education for sure.

SPEAKER_00 25:15

Very well said. And on that point, then I want to dig into the specifics, but I think I'll just start with more of a general question. How are you seeing generative AI being used and rolled out in K-12 education now, you know, more than three years on from the release of ChatGPT?

SPEAKER_01 25:31

Okay, so at the schools using it with kids level, there's a lot of chatbots, tutoring. If you go to the ISTE conference, I don't recommend that you do. And I've been there and I've actually presented there. I've presented there for a vendor there, was not a generative AI vendor. Actually, I did do one. I did a session about uh the Eliza effect and historical speakers. It was for a teacher-centered generative AI. Like they brought me in to do it, it was pretty cool, but their stuff was not student-facing, so I don't think they thought of it as a threat in any way. But ISTE is now overrun with just AI apps, right? And that's a conference for teachers or that's a that's a conference, the International Society for Technology and Education. So that's for it's supposed to be, you know, K-12, it's very expensive. So why it winds up being mostly like district folks who can expense it. So a lot of it is some sort of chatbot teacher, tutor, and then uh lesson plan generator. And and as a lesson plan generator, that just breaks my heart. I think about the best lessons I've ever done, and they've involved one million small decisions over time, whether it be my summer or like, you know, and I get it, teachers are overworked. They have so much work to do. And I don't know that large language model chatbots are the answer to that. But that's the big thing. Now, on the student side, there are some that, you know, interact with students, but a lot of students will turn to generative AI, whether, you know, usually Chat GPT of all things, to get answers. And I I personally think that it's a matter of solidarity with students. If our students are so overworked, if they feel so overworked, that we need to like reconsider a lot of the systems, a lot of the assessment, a lot of things like that. I think that's the answer to that. That's not easy, that's not a quick fix. But students relying on a chat bot, which we know has an errant inaccuracy. We know, I don't know if it's like documented by research, but it's documented anecdotally a ton. That long-term use of chatbots, at least in some people, has some bad effects. Um, so there's some things that we need to think about as far as okay, if we know our students are using them, why? And what, like, what is it that we can do to get students to a place where curiosity uh is driving the learning, not I have to get this done is driving the learning.

SPEAKER_00 27:53

Yeah, I think that makes a lot of sense. And I I want to dig in on some of the different places where we potentially see chatbots, right? Or, you know, generative AI being used. And I wanted to start with the students because I feel like, you know, we've had a lot of stories, you know, a lot of media reporting and things like that about student reliance on chatbots, the the amount that students are using chatbots. And I think some of that has been kind of debated, the the degree to which that is happening or like how how they're using it. Um I remember one time seeing like the user graphs of Chat GPT and like when the school year would end, you would see the user numbers drop off, which you know kind of suggested a lot of students were using it. But yeah, so I wonder what you're seeing in the kind of student use of generative AI. And I imagine that is probably more prevalent on the high school end where you are rather than the earlier grades. But yeah, what are you seeing?

SPEAKER_01 28:48

So the big thing I'm seeing is students turning to it to help them understand uh more difficult concepts. So I'll give you an example. So I was teaching a course that had some very difficult reading, some very dense, difficult reading. And so I had a student who was turning to it to kind of summarize it. I maybe they were, they I think they were taking like the big questions and themes from the reading and entering it into ChatGPT and seeing what came up. And what happens is, and this happens to a lot of people seemingly who use generative A on chatbots, they were very confident that they knew uh what was in the text. And then when I started asking a few questions, it was kind of oh uh, you know, like so I think that's a big thing that students are using it for, trying to kind of like figure things out. And I think a big problem that I think also, you know, the foundational knowledge about it should be stochastic parrots, Eliza, but also just like little language uh that gets away from anthropomorphization. So verbs like ask, that's what they would say, that's what most people would say, but we don't ask chatbots things, we ask people things. So it so I would say they enter text in with questions about American history concepts, about a chapter in a textbook, and then the chatbot generates text that in their mind, to their mind, responds to that. I know that's very difficult, but otherwise we're just anthropomorph anthropomorphizing these chatbots, and that cause causes all sorts of problems. But anyway, so that's what's going on there. Um, that's what I'm seeing as far as student use. And I think it varies. It could be, I've seen, you know, I saw a student, you know, they copy pasted an answer to a question, clearly from Chat GPT. And so I just had to say, like, look, you know, the policy of this was no generative AI allowed. Um, as a teacher, I want to hear from my students. Like, what an LLM generates is of no interest to me. One, it's just not interesting. And two, it doesn't help me assess a student in any way, right? Like if it if an LLM generate, so what does that tell me about a student's understanding of the con of the subject?

SPEAKER_00 30:59

I have a couple questions based on that. The the first is does discussion of the use of generative AI or or even like what it is, does that kind of come up in conversation in the classroom? Is that something that you discuss with students at at any point?

SPEAKER_01 31:15

So yes and no. As I've come back, I've come back to the classroom. So I came back in 2025 and I'm covering for another teacher. Uh next year I'll be able to do it all on my own. I'm very excited. And so then I'll I'll cross that bridge. Um, and I do think I have some ideas, I can get into those about how I might do that. Um, but I'll another example. So I was talking to a student about a rubric that's generated by a third party, long story short, it's like these advanced exams. And so there's this third-party rubric that we just as teachers use. And the student cited, he he said, Chat GPT set, and I said, That's not a source, right? Like, if you show me the document, the PDF or the web link, yeah, great, we'll have that conversation, but ChatGPT is not a source, and so that's it does come up. Um, I've had students say that, like, oh yeah, I love chat, and they they sometimes call it chat like it's a friend, and that's kind of I've heard that as well, yeah. Yeah, um, again, if you're told this stuff is like going to be sentient, and because chatbots use personal pronouns, which is a choice that developers make, these are choices that are made. So I understand why that happens. And you know, I was a teenager too. And if you told me that there was a shortcut to getting something done uh in most of my classes, in my history class, I wouldn't have because I was just interested in history. But to get through my AP bio class, yeah, I would have, I would have definitely done it. So I you have to understand that temptation as well. But yeah, I would have done it. It wouldn't have really helped long term, but I would have done it for sure.

SPEAKER_00 32:55

Yeah. I'm sure I would have been doing the same as a high school student. Like, you know, there's there's no question about it. I wonder as well, like I've heard stories of teachers and and even university professors also changing the way that they do like assessments and tests now that generative AI is a thing. I wonder, has that entered your kind of thinking? You know, again, I know that you're covering this year. I don't know if you can change the way that works, but if you're thinking about how you might approach things next year as well, like, you know, is there less kind of like essay writing at home and more writing essays in class so you know that, you know, that that LLMs can't be used while doing it? Like, I don't know, is that entering into your thought process at all?

SPEAKER_01 33:32

That definitely is entering into my thought process. I do think, and I think it's just more equitable in the first place that bigger assignments be done in classroom anyway, because you you just don't know the obligations of a teenager when they're at home, and that's very socioeconomic driven. So that's just more equitable in the first place. And I will say I don't I'm not in love with the idea of if a student has a long piece of writing, they have to do it paper-pencil because there are some students who use speech to text. There are some people for who for writing, it's a challenge. So I don't love that. But if it's in if it's taking place in my classroom, then I have more eyes on it and I have uh more ability to do that. But having said that, being more analog is definitely, oh my goodness, it's time to rediscover that. I recently inherited a project from a teacher I'm currently covering for in history class called the Stamp Project, where students would just draw and design a stamp, like a postage stamp of any big topic in American history, and it can be anything. And the results are amazing. I'm taking pictures of them and texting them to my friends, just saying, like, can you believe what these kids did? Um, I texted them to the teacher I'm covering for, and and the teacher I'm covering for almost expected it. They almost were like, Yeah, I know these kids are amazing, you know. So um that to me, and and that class is at a higher level, and I'm saying to myself, if I'm teaching that same content at a lower level, we're still doing that project next year. Like that's that's happening. Um, because it is so wonderful to see that art. Um, so I think that's that's gonna be a big driving thing is like what can we do analog? What can we do in the classroom? And knowing that if I give my students a very big piece of writing, then it's gonna have to be, you know, maybe the the brainstorming, the uh the first pieces in the classroom are gonna be maybe more what I focus on when I assess it. Um, because I know that there will be a temptation there for sure.

SPEAKER_00 35:37

Yeah, uh I think that makes a lot of sense. I I was wondering as well on that front, like, you know, as as you mentioned earlier, we do see these reports that uh, you know, extended use or or using uh and relying too much on chatbots can affect people's critical thinking skills, their cognitive abilities, their ability to retain information. Um, do you see much of that or do you worry about that kind of in class with the use of these technologies?

SPEAKER_01 36:01

Oh, I definitely do. I the cognitive offloading, and I think just the overconfidence, there's now there is not a humongous research base here, but I'll give you an example. One of the few ones that exists, it's a, I think it's a pen study, and I'm gonna do my best to summarize it. But basically, you know, students were given a math problem, and then they were given an AI tutor, and some weren't. And what happens is in the immediate aftermath, students do great. Okay. And then on the long-term assessment, not only do they think they did well, but they don't do as well. So they have this overconfidence and underperformance. And I can really see that that makes a lot of sense to me just from the little I know about LLMs and what I know about teaching and students. It yeah, yeah, that's gonna be it. That's gonna be a real challenge for teachers, is students who are overconfident because they've interacted with a large language model, and then, oh, but you didn't really have real learning experiences where you got your fingers in the dough. So, yeah, that might be a real uh thing to consider as we go forward.

SPEAKER_00 37:08

I I don't know about you, but as I hear you describe that, what comes to mind is kind of the discussions that we were having a few years ago about students, you know, being at home and and learning virtually during the pandemic and how that potentially impacted social skills or or learning and all those sorts of things. And it felt like at the time there was a lot of concern about that, right? You know, obviously some of that was motivated by people who didn't want lockdowns to happen and and all that kind of stuff. Um, but there were, I think, legitimate concerns there as well, right? Like it is important to be around people, to be, you know, in person when you're learning and those sorts of things. Um, and I and I I don't know, it's a bit weird to me, maybe that I feel like you know, those kind of concerns about uh students and their learning experience and kind of their mental development and all those things are maybe not getting that same degree of attention as I feel like virtual learning had a few years ago, you know, and I'm sure that there are like commercial reasons for that and all sorts of things. But yeah, I wonder if if that's something that comes up for you.

SPEAKER_01 38:12

Okay. So personalized learning, the idea is that each student has different lesson materials or lesson materials that meet their unique needs andor interests. And especially if we go to say, really the whole decade of the 2010s, personalized learning was just everywhere. And technology makes personalized learning so much easier. And what I've noticed is that in the 2020s, personalized learning has been dropped. It has been dropped like a hot potato. And I my guess on that is that it's because as a society, it's becoming very obvious that we don't have a collective agreement and understanding on truth. And so the idea of things being more individualized is actually not something that people that people really want to see. And of course, an LLM tutor is the is like personalized learning on technology enhanced steroids. Now, again, steroids do actually make you lift more and hit home runs farther, right? Um, LLMs, sometimes uh anyway, what I'm seeing there is that the idea that each student is in front of a device, which is now getting so much pushback in the United States of America, so much pushback from families. Um, but that's really where this is where this would be going. If we're saying that we want students interacting with a Chat GPT or other LLM, that's where it's going. We're going to like literally student and device one-on-one. And as far as community, collective understanding of the truth, social skills, that's not going to help that situation at all. Um, and I I actually think I'm really almost surprised at how excited people in K-12 are about LLMs and generative AI tutors when personalized learning has become kind of passe. I'm surprised that how that that's happened because it's kind of like that is personalized. What do you, you know?

SPEAKER_00 40:09

So yeah, that's fascinating. That it's really interesting to hear. I wanted to flip it a bit. Um, you know, we've been talking about students, what you're seeing with the students' use of AI and those sorts of things. If we turn it around to the teacher side of things, obviously, you know, teachers are using generative AI as well. You know, you mentioned some uses earlier. I wonder what you're seeing on that side of things, on how teachers generally are thinking about generative AI. You know, is there a lot of kind of critical uh views on it, or are people adopting it as you say, because they're overworked and they see this as a potential tool? And do you see the adoption coming largely from the teachers themselves, or is it often being pushed by administration, by the school board, things like that?

SPEAKER_01 40:54

So what I'm seeing is a huge it's a huge mix. And I I I'm happy I don't think that you can say by and large teachers are doing A or B. I think it's a mix. And I think one, it's definitely not informed by deep understandings of understandings of large language models. And when I say that, something I hear a lot is that, oh, well, students and teachers need to learn how to use these tools. But using them is so easy. You literally enter a prompt and then something comes out, right? Using them is so easy. Um, and if we think, well, you have to refine your prompt, well, then maybe it's not like then then come on, like if you have to do if you have to refine things 50 times, is there's really a time saver. But what's happening is you see a lot of of things that are true about teachers. So teachers are overworked, teachers are also huge BS detectors, like teachers are so skeptical. You know, they're kind of trained to be like that. Life experience makes them like that. And so you have the ones who they've never heard of the stochastics parrots paper, but they're sitting there saying, like, come on. And they've also been through ed tech overpromising over the years. So they're kind of like, come on, really. And then there's the teachers who are like, look, you know, did you I just want to generate this rubric and be done with it? And then it gets generated, right? I see some that just ignore it, just don't care one way or the other. I think a lot of teachers are annoyed at students for using it. I think the student usage thing is a lot more complex thing that we have to, again, have some solidarity with students around. But I think it's all over the place. And I would like I just want the conversation to start with the actual experts, with the computational linguists, with with Eliza and with Weissenbaum. Like, let's start there. That's and and that's just not part of the conversation right now. That's what I want to see happen.

SPEAKER_00 42:51

That's fascinating to hear. And like, I'm not, I'm not surprised, right, that there's going to be different approaches. And obviously, that teachers are going to approach this with a critical eye because you know they've seen so many of these promises before, not pay off. I did want to go back to the part about like administration and school board. Like, is there trainings on generative AI that are put on for teachers? Is there kind of like a push from higher up to use these technologies?

SPEAKER_01 43:17

That really varies school by school. And you see. So what I'm seeing is the big thing that gets people in positions of power to pause is the environmental stuff. For whatever reason, the environmental thing, now there are a lot of other documented harms of generative AI. When people hear that environmental thing, they pause. Some people mention it and then just go the heck on and keep going forward. Um, other people, though, say, like, hey, this is really a problem. Because now we know about data centers. Uh, we know, well, at least I I think a lot of people know about Boxtown, where because of Elon Musk's data center, people Saying how how come I can't breathe? And we know, by the way, that this isn't like a you and me, Paris and Tom thing. This is a historically marginalized and oppressed people environmental racism thing. That's what generative AI is. They're not building data centers in the affluent places. But so we know about that. Um, but then there's the other people who I think they have that fear of missing out. They do not want to be considered archaic, they do not want to be considered out of date. They are grappling with the idea that K-12 education is job training. And so therefore, we have to have our students current and relevant, uh, and our instruction current and relevant. And so I think that's what's happening there. And just like it's really hard to say that all teachers are in or out. It's really hard to say whether or not all school boards, all administrations are in or out one way or another. It really runs a gamut. Um, but I do think they all feel that pressure to like, hey, we're not outdated, we're not dinosaurs, we are cutting edge. Um, but I I also I kind of wonder, you know, it's becoming such a thing that there's a backlash against generative AI, against tech companies, against it being mentioned in commencement addresses. But I wonder if that whole cutting edge thing is going to kind of uh diminish a little bit. I wonder if that's almost going to be kind of like, oh, you know, we don't want to, you know, like traditions, and again, I get it. Tradition in a lot of ways is traditional things that harm people. And so tradition isn't a really a great value either. I like to say it should be about solidarity with students and teachers. Like, what is the most solidarity place to be for those groups? And that's what should animate it, not innovation. And I will say, by the way, if there's been a word in K-12 in the last 10 years that has done more harm, I don't know of any more word or any word other than innovation. What's happened is what little Johnny and Hakeem and Aaron and Sandra need has been put on the backburner to innovation. Innovation at all costs. And innovation at all costs is not what our kids need.

SPEAKER_00 46:23

It feels like that's not just a K-12 schools thing. It feels like it's our whole society, right? Everything has been put on the backburner to so-called innovation, uh, whether it is even that at all, right? Um it that that's really interesting to hear. Um, and and I wonder, you know, as we start to kind of close off our conversation, what you see in the pushback to this, right? You know, you you mentioned the data centers there, and obviously we're seeing a growing pushback to data centers. And I feel like we're seeing more and more reporting on the use of generative AI in schools. So I wonder what you're seeing on your end to the rollout of these technologies and whether students are pushing back on them, whether teachers are pushing back, how how parents are reacting. How is the discussion evolved in schools around uh generative AI?

SPEAKER_01 47:13

So in the Philadelphia suburbs, there's a district where I think there are up to like 600 parents who are petitioning that their kids no longer receive a Chromebook. Um, they actually have a they have a motto, and I see it, they put it on lawn signs, like it's an election campaign. It's called pencils up, screen down. And that I think is going to be the biggest factor. So there's been some resistance from districts who say, hey, we need one-to-one devices in order to do education. And I've thought about that and I realize like, I do I need technology to do education? Yes, like I need my computer, I need a projector, I need a photocopier. I absolutely need those. But if you were to tell me that if I have those and my kids lost their devices tomorrow, I feel like I could deliver education. And if I think saying that you couldn't is a very dangerous thing. I get it. It's it's I'm not very good at organizing stacks of paper. Uh, you know, I get it. Uh you know, I oh, Mr. Mullaney, what did we do yesterday? And then I have to go find it because I'm I'm not great with that. But the idea that we can't do school without screens, and I I get it, I I've already counted a screen for myself and a projective screen, so I've already I'm not at zero screens, but I don't want to be in a position where I say we have to have student one-to-one. I think that digital instruction can happen, but it can be, you know, maybe a little bit more special. And it should be things that really excite the kids. There's an app that I plan on getting into this summer. Um, ArcGIS has these story maps that you can create. And I'm like, oh, I want to play with that. But do we have to word process every single day? You know, like I think we can have some fun with it and do some interesting creative things, but for it to be a daily driver, I think we need to like really have some conversations about that and really have some thought about that. But so that's what I'm getting at is the backlash against screen time, which I get it, it doesn't uh in some ways it doesn't address inequity, it doesn't address a lot of the real, real big problems with education in the United States, segregation, which is still very much a big problem, inequality, wealth inequality, all that stuff. Um, but that backlash against screen time, I think is is really starting to take hold. And I think it's only going to get bigger and better. And I think in a lot of ways, the promotion of ed tech and innovation over all other things, it's kind of like, hey, you know, you were cruising for a bruisin and a bruisin is common. And, you know, I think it's like kind of justified.

SPEAKER_00 49:54

I love that. A bruisin is coming. They they deserve it 100%. Um it's it's it's fascinating too, right? Because it feels like at the moment when they are really kind of even pushing to ramp up the amount of technology, you know, to get people using generative AI and all this kind of stuff. It's like you have this counterforce of parents, you know, and I'm sure of students and teachers to a certain degree as well, who are like, hold on, it's not just the generative AI that's the problem here, but maybe we've gone far too much into this adoption of screens. Like, I don't know if your school, I know it's becoming really common in Canada now for phones just to be kind of banned in schools or like they need to stay in the locker kind of a thing through through the day. That's becoming really common here. Yeah, I'm I'm not surprised to see that there is this growing, not just pushback to, but skepticism of you know, the idea that everyone having a screen, everyone having a Chromebook or a laptop is the best way for people to be learning.

SPEAKER_01 50:54

Oh, absolutely. And Pennsylvania doesn't have a law about cell phones. I think New Jersey does. I think New Jersey, right next door, has a ban. And yeah, cell phones are that's just we all know when we have a cell phone near us, we're distracted. That's just basic common sense, right? That's just what is what that is. But I do think a lot of people are starting to really question and try to be a little bit more analog for sure. I'm really curious to see how this plays out, but we are definitely at a moment of backlash. I I yeah, I'm I'm kind of curious how this plays out. But I my thing is I just want some critical thinking. I just want like if if I can attend a professional development session on generative AI and they acknowledge that the stochastic paper exists, they mention Eliza and what what and Weissenbaum's conclusions from Eliza, then I'm like, okay, we're at least thinking about this critically. That's what I want to see. I want to see that in K-12, and I just haven't seen that in that discourse yet.

SPEAKER_00 51:57

Yeah, if we're gonna have the discussion, let's at least have the broader discussion, not just the kind of booster-y, hyped-up discussion that tries to get us to adopt it.

SPEAKER_01 52:07

Exactly right. And let's also keep our examples relevant. Robotaxis are not part of the decision about whether or not large language models are part of effective pedagogy. They're not generative AI. All right. As terrible as it is that technology labeled artificial intelligence AI, a marketing term, is used in determining who to bomb, as awful as that is, I don't think that bears one way or another about whether or not a large language model should be used in a K-12 classroom. Right. So let's keep it to generative AI, to things that might actually be used in the classroom, and let's be fact-based, not boosterism, and let's not quote people who are selling it to us. Let's definitely not do that.

SPEAKER_00 52:52

Oh my God, that's such a problem. I I've done I've done interviews before where they push back on me for being critical of the technology, you know, saying, oh, you just don't like it because you're more on the left, all this kind of stuff. And and I'm like, when you have the promoters on here, do you push back on them and be like, oh, it's because you're financially incentivized to push us to adopt it? No, they don't, right?

SPEAKER_01 53:14

Right. And I I would say I don't even think that I mean I get it. Like the Trump administration loves generative AI. Like that's that's a hundred percent true, um, which is funny because I see a lot of people who I see promoting it for K-12, if you were to look at their other posts, they're very anti-Trump, but they don't make that connection. I just think though, as far as the backlash, as far as the backlash, especially in screens, but also there's the backlash going on in this country about generative AI and data centers, that cuts across political uh spectrum. That's that's Mondami, and that's Trump, uh, as far as who who the supporters who are very, very concerned about these things for sure.

SPEAKER_00 53:58

Yeah, absolutely. Parents uh, you know, of either political persuasion are on this. Um, Tom, this has been really fascinating uh to learn about what is going on in K-12 schools, what you have been seeing, um, and you know, the real important questions that we need to be asking about this technology and and what we're doing to kids, but also to our broader society in rolling this stuff out in schools and not thinking about the bigger questions. So I really appreciate you taking the time to come on the show. Absolutely. Thank you so much. It was a pleasure, Paris. Tom Mullaney is a high school teacher in Philadelphia. Tech Won's Save Us is made in partnership with The Nation magazine that is hosted by me, Paris Marks. Production is by Kyle Houston. Tech Won's Save Us relies on the support of listeners like you to keep providing critical perspectives on the tech industry. You can join hundreds of other supporters by going to patreon.com slash tech won't save us and making a pledge of your own. Thanks for listening and make sure to come back next week.