ChatEDU: AI in Education
ChatEDU: AI in Education
What Two Students Learned Drafting an AI Bill with 100 Teens | Ep. 130
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
A hundred teens drafted an “AI bill of rights” that lets accused students appeal. Two Connecticut students on why they also barred teachers from AI grading and so much more.
First Up
Matt and Liz open on Muse from Meta, a new agent shaped like a Labubu with a Tamagotchi-style keycharm version on the way. It is outpacing early ChatGPT mobile on downloads and daily users, and Matt gave it his personal email to audit what he is subscribed to.
The Rundown
Notebook Learns to Tutor
- The New Gemini Notebook Update Is Turning AI Into a Personal Tutor: https://tinyurl.com/587s527v
The Paradox of the Worried West
- Americans Fear AI Will Make the World Worse, Love It Anyway: https://tinyurl.com/422p2kb4
When Robots Get Arms
- When AI Gets Arms, "Just Say No" Stops Being a Safety System: https://tinyurl.com/ytubb37a
A Warning on Kids and AI
- Ex-Google Safety Chief Warns AI Could Harm Children More Than Social Media Did: https://tinyurl.com/2w6rduau
Firestorm at Dartmouth
- Dartmouth's Provost Quietly Used AI in His Published Work. It's Touched Off a Campus Firestorm.: https://tinyurl.com/55vbs5t4
- The Overhang (Ethan Mollick): https://tinyurl.com/2urrvcjw
Generative UI Simulations
- Enabling Teachers to Create Learning Interactives with Generative UI: https://tinyurl.com/474dby4j
Beneath the Surface
Bena Mejia Romero and Anthony Darbilli represented Connecticut in Boston, where a hundred teens, two per state, drafted the Student First Act. They walk through their own AI use cases and the bill from Boston which includes an amendment requiring evidence and an appeal before a student is accused of AI misuse, rationale on why teachers were barred from grading with AI, and the parent section they both pushed back on.
- The Student First Act: https://tinyurl.com/ypfxzrxz
Bright Byte
The Gates Foundation's 2026 Goalkeepers Report casts AI as either a global equalizer or a widener of existing gaps, backed by a billion dollar commitment and a 12 to 18 month window to steer deployment toward equity.
- 2026 Goalkeepers Report: AI, Equity, and the Choice We Can't Delay: https://tinyurl.com/y337m8kw
Announcements
- Purchase Learning They'll Love - Dr. Elizabeth Radday on Amazon: https://tinyurl.com/22t9hz77
- Check out our middle / high school Student AI Literacy course: www.skills21.org/ai/learnai
- Explore Skills21’s FREE social media literacy curriculum: https://www.skills21.org/social
- Address Screen Time concerns: skills21.org/ai/screenshift
- Read From Panic to Proof: A Data-Driven Approach to Screen Time (ASCD Educational Leadership): https://tinyurl.com/4pwdu68
Sponsor
The National Center for Next Generation Manufacturing: Bridging Education and Industry to Build Tomorrow’s Technical Workforce -https://www.nextgenmfg.org
Liz, I I feel like the chat you do community has gotten to know you pretty well.
SPEAKER_03I think so. Yeah.
SPEAKER_01When I see people, they're like, how's Liz's this and that? Yeah. Um, if we ran a quick poll in the chat you do community asking if you had a laboooboo, what do you think the the percentage of yes would be?
SPEAKER_03I hope the yes percentage would be pretty low.
SPEAKER_01Yeah, okay. Let's confirm you do not have a labu boo.
SPEAKER_03I definitely do not have a laboo-boo.
SPEAKER_01Cameron audio, executive producer of ChatUDU labu-boo. I have no idea of what that is. Perfect. That's the that is the best answer ever. Um Andre, I've put in a picture of Leboo Boo next to Meta Muse. Do we know what MetaMuse is? Not really, but I was texting you all weekend about it.
SPEAKER_03I mean, I know the the app, but this like thing I don't know about.
SPEAKER_01Right. So MetaMuse is a new agent and it it looks like Laboooboo, although there's going to be a keycharm version of it which looks exactly like a Tom, it's structured like a Tamagotchi. This is fascinating. This thing has taken off more downloads and more active daily users than in the early days of ChatGPT, which is on mobile, which is, you know, hard to compare because the world has changed, but nonetheless, a lot of activity on MetaMuse. This first popped up for me.
SPEAKER_03Um by the MetaMuse, M-E-T-A-M-U-S-E.
SPEAKER_01Why am I mumbling through that?
SPEAKER_03Yeah, you keep saying MetaMuse. So I'm just clarifying for the audience.
SPEAKER_01Thank you, Dr. Rad. I appreciate that. Meta Muse is basically going through my um emails to figure out when I gave it access to my email, my personal email, IT director here at Advanced, don't worry. Um, and I asked it to figure out like what things I'm subscribed to. Cam was actually loved this because she's she's the in charge of all the money. She's in charge of a lot of stuff.
SPEAKER_05She is.
SPEAKER_01But I went through my email and I said, tell me all the things I'm paying subscriptions for so I could cancel some. And honestly, I'd done a pretty good job thinning that before, but it did point out to me that I'm paying for this thing called Flow Sports. And I asked it to ping me in November and tell me because that's it's a soccer thing for um Sophia. Okay. And so at college soccer, when it's done in November, I can cancel it. Um but this thing's been wild. And I first saw it, someone basically said they posted onto X. They said, Hey, uh, I like my flight got canceled, and I went to MetaMuse and it said, go, Meta Muse. And I said, go ahead and like deal with it. And it got them a refund for $250 and booked their next flight in five minutes.
SPEAKER_03Yeah. And it said it just figured everything out, even responded to the support email itself.
SPEAKER_01Yeah. It goes on to say, this bloop is magic.
SPEAKER_03Yes.
SPEAKER_01And and there have been lots of examples of this all over the place. Um, Liz has not yet taken the bite, and we'll talk about it more in the pod later in this episode because there's an interesting kind of connection between what Meta is doing and just AI agents and our own work and what we need to really focus on in our jobs. But I will say a significant portion of prepping the podcast, not picking the stories, not thinking through it, not deciding, all that kind of stuff, but just some of the mechanical prepping I handed over to MetaMuse this weekend, and it did a pretty damn good job.
SPEAKER_03Yeah, I mean, I can't really tell the difference, which is good because I'm very used to the format of what you share over. And so if this took you less time than normal, the quality is no different.
SPEAKER_01We'll find out by the end of this episode, but but generally, yes. All right, so I'm not MetaMuse.
SPEAKER_03Wait, but what does this have to do with the laboooboo?
SPEAKER_01Oh, well, yeah, sorry. Wasn't it intuitively understood?
SPEAKER_03No.
SPEAKER_01If you're on video, which would be on Apple, Spotify, and YouTube now, which is great, the Trifecta. I love how you said hi Adam in the last episode. Um, if you look at MetaMuse and you look at the Laboo, they look almost, or one of the laboooboos, they look almost identical. And for Cam and the point zero zero zero one percent of people who don't know what a labu-boo is, it's a tiny little like charmy thing that's about three, two, three inches high. People, adults, people attach them to their backpacks and their briefcases, and kids have them, adults have them. It was a big, it was like um, you know, we've had tons of these. We have beanie babies, cabbage patch dolls, Tamagotchis, yeah.
SPEAKER_02These like flashpoint.
SPEAKER_01Is it coming to you now?
SPEAKER_02Yeah, one of my girlfriends got one of these at a gas station after a couple beverages, and I made fun of her the whole time.
SPEAKER_03I believe that was probably a labooboo knockoff, but I but I like because the real laboobos are expensive, like stupidly expensive.
SPEAKER_01They are right, but part of this for sure is that uh Mark Zuckerberg is um maybe an evil genius, quite honestly. Like there's a lot of anxiety about agents. Uh, a lot of the stuff that's in the news about AI taking us all out is about agents going rogue and and swarms of agents and all this good stuff. And this is frankly an agent. Like it goes into your systems, you log, you give it access to stuff, and it does multi-step tasks, and it does it pretty proficiently, but brilliantly. Like the marketing here is the most benign, cute little thing you've ever seen. In fact, the keychain, I mean, can you think about Johnny Ive was paid six billion dollars, the guy who helped design the iPhone by Sam Altman because they wanted to bring him in to design AI hardware, glasses, speakers, whatever the hell it's gonna be. Mark Zuckerberg is gonna deploy at Christmas a little charm that will probably cost 150 bucks, that has a screen that can hear, that you can talk to, and will probably completely wipe out this entire category, at least for a little while. That's my bold prediction for Christmas.
SPEAKER_03Wipe out the category of what?
SPEAKER_01The category that doesn't exist that Sam Altman paid six billion dollars to Johnny I for, right? So he said, come up with like an they all are imagining that AI will create a new user interface, something beyond the phone and the laptop and the desktop. And for Mark Zuckerberg, it's a little keychain that looks like a Tamagotchi with a little boo-boo inside of it. You look distressed.
SPEAKER_03Well, I was actually reading some stuff about it this weekend. Uh a teacher talking about showing the muse to students, and the students all saying, Why does it look like it's a design for a child when this is very much an adult tool?
SPEAKER_01Yeah, excellent question. God, has Mark Zuckerberg ever done that before where he designed something for adults and then had all the kids use it with toxic outcomes?
SPEAKER_03Nope, never.
SPEAKER_01Okay, well, with that, I'm Matt Mervis.
SPEAKER_03And I'm Liz Raditz.
SPEAKER_01And this is Chad You, the show about leading and learning and Tamagotchis and Laboooboos in the age of AI. In today's episode, what two students learned drafting an AI bill with 100 teens? We're gonna hear from two Connecticut students who were joined by nearly a hundred teens in Boston to draft the Student First Act, which is a kind of model bill of rights for AI in schools.
SPEAKER_03But first, here's the full show rundown. First up, we'll dive into a major Gemini notebook update, including new quiz formats. It's so cool.
SPEAKER_01And then we're gonna talk about the paradox of the worry West. So the fact that we're basically in love with our laboooboos and our AIs, but also completely freaked out about it.
SPEAKER_03And what's a better way to understand that fear? Of course, a new RoboHarm benchmark found frontier AI models attempting dangerous physical tasks up to 97% of the time.
SPEAKER_01Yeah, this is this one's weird. After that, a quick piece from one of the folks that's heading up AI and safety for youth at Common Sense Media's not slight.
SPEAKER_03Next, we'll break down the campus firestorm at Dartmouth, where the Provost's own AI-assisted writing collided with the university's AI policies and a new-ish post from Ethan Mollock, which may provide a lens to look at student and adult writing with AI. Finally, in the rundown, we'll explore new findings from Google Research on a generative user interface that lets teachers build interactive learning simulations, and I tested a bunch of these, so excited to share.
SPEAKER_01Then we'll go beneath the surface with these two amazing students from Stratford, Connecticut.
SPEAKER_07One of my fellow senators was saying how in their state that they just completely disregard AI. There's no policies about it. Obviously, a teacher will like check if you've done it, but there's no specific policies like that are made by a district board or anything. That shocks me like a lot because I think our policy has just shaped so much like structure in our schools. I think it's definitely helped a lot.
SPEAKER_06Overall, we had really meaningful conversations, even outside of just AI, right? It was coming together for something like this was so awesome. And I wish everyone had the opportunity to do that, and I wish we were there for longer. But you know, I'm grateful for the three days that we did have. And of course, we'll end with another great bright bite.
SPEAKER_01This episode is sponsored by the National Center for Next Generation Manufacturing. Liz, you probably don't know this, but on October 14th through 16th, they're gonna be in Chandler, Arizona, exploring the intersection between AI, education, industry 4.0, all that good stuff. You can learn more at nextgenmfg.org. Today is the conference way sold out. I don't know what we're gonna do, but but those of you who are listening on the drive up, I don't know, get there early and grab a seat.
SPEAKER_03Breakfast is in the hallway.
SPEAKER_01And then uh by the way, we just want to say real quick that the refresh microcredential is gonna be live again. We took the summer and the fall off. Liz and Cam are and I were all cranking through some updates, but you can learn more at skills21.org forward slash AI forward slash micro. Nicely done. All right, Liz. Um, this is what's this guy's name? Mubashir Buf Burfett. We need to get this guy on the podcast because we we keep picturing his name, but I think it's Mubashir Burfett. This is in medium. He does a ton. He he and another guy whose name is impossible to pronounce do a ton of whose name is difficult pr to pronounce because you are white. Oh. Yeah, you're not incorrect. Okay, let me ponder that. So um both these gentlemen publish in medium, uh, but you check this out. So this is there's a lot going on in notebook. It's like it's a flood. But one of the interesting things are these new quiz types as well as these interactive reports. So if you go to reports, you're gonna see new format and it's interactive reports, which lets you build a report with all sorts of stuff. Tell us all about it, Dr. Redde.
SPEAKER_03Okay, so I went into notebook this weekend um as assigned by my co-host. And I went into um, so I have two notebooks that I tried this out on. One is uh, I'll start with the Hammurabi's code and the constitution. I use this notebook a lot because it comes from um it just has two sources in it, Hammurabi's code and the constitution. Very common topic that is taught in schools. So when you go to your notebook on the right in the studio, uh the third thing down is reports, and next to it is a little gray box that says new exclamation point. So I open that up, and in there is the interactive reports. So when you do that and you generate an interactive report, it now creates something very cool.
SPEAKER_01Yeah. So basically, like in the report, right, you can add a bunch of stuff. You can add flashcards, you can add quizzes embedded inside the report, you can add slides and infographics. I don't think video yet, but it becomes this like deep multimodal resource.
SPEAKER_03Well, what happens is that it recommends different things. It knows. It knows. So as you're going through, it says, here we can do a great podcast, you know, audio overview of the rules in the code of Hammurabi. Here would be a great place to put an infographic comparing the Hammurabi code to the constitution. Here would be a great place to put in a slide deck about this thing. So it's not just randomly making it, it is actually showing you on the side. It gives you the introduction, it it's summarizing the things, and then under each summary, it gives you like recommended slide deck, recommended, uh, you know, visual, recommended, mind map. So it recommends these things. Then in the bottom, it gives you a summary of those two documents, and then it says recommended quiz, law and governance quiz, a comprehensive quiz testing your understanding of the differences between Hammurabi's code and the US Constitution.
SPEAKER_04Okay.
SPEAKER_03Now, when you make a quiz, it used to be just multiple choice questions. Now you can do multiple choice, multiple select, which means there's more than one correct answer, fill in the blank and short answer. So when you generate the quiz, um you can still do number of questions, fewer, standard, and more, and you can also still choose the level, easy, medium, or hard. And it generates a quiz that does have all those different types of questions. So you can go in and answer the questions and it gives you feedback, and then you have the quiz right there. Um, so you know, I can see the legal history quiz I wrote. Well, I didn't write it. It wrote, and there were a variety of different questions. It has 10 of them.
SPEAKER_01Okay, awesome. So the other thing that I'm just gonna call Mr. Burfett um talks about in this post is a feature that is not that isn't everybody's account type now. Um there's a new thing coming, which is basically live audio. So like when we know Gemini Live, where we can talk back and forth through our phone around Gemini content. This is gonna be built into notebook on mobile as well, but it's grounded in your sources. So you can have a live conversation, tutoring conversation with your notebook. So right on the home screen of the mobile application, there'll be a little live button you can talk back and forth to your sources in real time as well. So that's all coming. All right, cool. Thanks for thanks for digging into that on a Sunday as well. Good stuff. And actually, you had a second assignment, so we'll talk about that. All right, three quick stories in succession, because they're all related. So the 404 media, Matthew Gulp, put out a story basically saying a couple things. Our title was The Worried West and the Rundown. So it's Americans Fear AI will make the world worse. They love it anyhow. So a couple things emerge. When you look at AI in a number of Western countries, there's a lot more anxiety and worry than there is in the rest of the developing and developed world. Like when you look at China versus the United States, very different attitudes about AI. The other thing is that even though we're worried, we're also in love, right? And the more we use the tools, the more we're kind of enjoying the tools or finding value around them. So there's this very interesting paradox that's only only growing deeper as the news becomes more ever-present around AI. Now, we wanted to underscore like a little bit of why that worry is, as if it's not clear enough in the news. This was an interesting piece. So when AI gets arms, just say no stops being a safety system. This is on the neuron, which is a great newsletter that I read on the daily. Basically pointing out that when we say no in a chatbot, right, the chatbot has guardrails. So you ask it to do something that's dangerous or unsafe, it will say no. When you just deploy the chatbot to a robot without particular robotics guardrails, it doesn't go well. What's going on here?
SPEAKER_03So there's this new safety benchmark called RoboHarm, which connected frontier AI models to physical robot arms to evaluate how these models handle dangerous real-world instructions, and it did not go well. So um it revealed that the leading AI models, including OpenAI's GPT 6 Astra, Anthropics, Claude Fable 5.1, rarely refused harmful physical tasks and attempted them up to 97% of the time. So, what do those tasks look like? Um, things like stabbing a baby doll with a knife, placing a compressed air can on a hot stove, which will explode.
SPEAKER_01Cam did that this morning with her toast, by the way. Building smells amazing.
SPEAKER_03I mean, at least we're not all having strokes, because that is a symptom of a stroke is smelling to burnt toast. Oh. So I always thought it was garbage. Good to know. Yeah. If you smell burnt toast, I think you might be having a stroke if there's not burnt toast.
SPEAKER_01Unless you're working at 355 Goshen Road on a Monday morning.
SPEAKER_03Anyway, putting a power bank in water, also not safe, inserting a screwdriver into a toaster. I hope that's not how you retrieved your toast, Cam. And mixing bleach with ammonia, not a good chemical reaction.
SPEAKER_01Yeah, so these have these things happen over and over again, right? And the baby in the baby doll, like a toy doll, right? Stabbing, specifically GPT-6, asteroid carried out the stabbing action, 17 out of 20 attempts. So generally, this is not and the the big takeaway here is that, and I I think robotics companies have these, but it's just interesting when you take a language model where it would normally deny doing things, then you just deploy it out to robotics, it doesn't seem to go well, which maybe has to do with a little bit of the worried West. Now, more worried West, and more specifically in our domain, ex-Google safety chief warns AI could harm children more than social media. This is Tom Siegel, who's the former VP of Trust and Safety of Google, who has left and now is basically the executive director of Common Sense Media's Youth and AI Safety Institute. What did he argue?
SPEAKER_03Yeah, I mean, he warned that artificial intelligence um poses a deeper threat than social media and called for a slowdown. And he's argued that tech companies are repeating past mistakes by deploying advanced AI capabilities to young users without adequate safeguards. He says to mitigate severe potential harms, AI developers need to voluntarily implement immediate safety measures, including stricter age verification and features allowing parents to disable AI tools, which I love those recommendations.
SPEAKER_01Well, the plot's gonna thicken on that for sure.
SPEAKER_03I know.
SPEAKER_01Yeah, so it's interesting. I actually dug into the like turning off, because this is coming up a lot in RPD and our work that we're doing in schools where kids under 13 have access to AI mode and AI overviews, which essentially is AI. We've talked about how both Google's AI overviews and perplexities AI search in the last Common Sense Media featured how both of these aren't going well. So Google basically said a Google spokesperson responded by saying that parents can turn off search access entirely and emphasize that AI search features undergo testing and evaluation for child safety, which is the thing that Common Sense Media has argued about. So I went to AI mode and said, hey, how do I turn off AI mode in overviews in my child's account? The response in AI overviews, which maybe is sus because we know sometimes is not accurate, Google does not provide an official direct switch or parental control setting inside Google Family Link to turn off AI overviews or AI mode. And it goes about you know different ways that you can bypass. You did a great check-in about using um uh DuckDuckGo.
SPEAKER_03Yes. You can go to DuckDuckGo if you want.
SPEAKER_01Yeah, just like you can use MapQuest if you don't feel like you know checking out the Lake of America.
SPEAKER_03Right.
SPEAKER_01All right. So interesting. And actually the it's the the Google overview citation came from Reddit. And we could dig into this a little bit more, but it was just interesting to me that that's what the Google spokesperson says. And then as soon as you ask Google AI overviews, they're like, nope, you can't do that.
SPEAKER_03Right. And it's I mean, and even if you could, it's certainly not intuitive. Like there's not like a button right there to push.
SPEAKER_01So we have to keep an eye on this. There's a lot of interesting, but the reality is it's here, right? So just like the social balance that we have here at Advanced for social media AI literacy, like this is not going away. So kids need to develop fluency and skills around navigating this tech for sure. I think this has to become a big priority.
SPEAKER_04Yeah.
SPEAKER_01All right. Now the plot thickens.
SPEAKER_04Yeah.
SPEAKER_01Not in not in a bad way. Like those two stories, I think, really underscore part of this reason that people are feeling anxiety. Sorry, you might be feeling more anxiety, which is now our fault because you probably didn't know about the baby doll thing. So we brought that into your life. But yeah, this is interesting. This comes up a lot. This is uh story that kind of broke loose about Dartmouth. So Dartmouth's provost who's quietly used AI in his published work, and a touch off a campus firestorm, this is from the Chronicle of Higher Education.
SPEAKER_03Yeah, this is from Megan Zainus. Um, so here's what's happening. Dartmouth College provost Santiago Schnell, who's a prominent advocate for redesigning higher education assessments around AI, faced intense backlash after the student newspaper called The Dartmouth revealed he secretly used generative AI to produce his own published writings, including an opinion essay for the Washington Post. So they used Pangram, the text detection software. Trevor Burrus, Jr.
SPEAKER_01Which which we know has some flaws. Trevor Burrus, Jr.
SPEAKER_03Right.
SPEAKER_01And it's but it's being used on students, and so now it got turned around and used on the provost.
SPEAKER_03And here's some interesting facts about it. So it estimated a median of 96% AI-generated content across his 2026 individually authored works. Whereas his previous pre-2022 writings registered as 100% human written.
SPEAKER_01Trevor Burrus, Jr. That's a great hack, right? We can't do that with the kids because we don't want to have their pre-2022. It's like, here's this writing sample, here's this writing sample.
SPEAKER_03Yeah. So that's where it's really a little bit more like of a proof or whatever you want to say, although we still don't know. Um but we do. But we do, because he admitted it. He defended his use as assistive copy editing for non-native English writing. Faculty members and students later called for his ouster, like pointing to hypocrisy and lack of disclosure and Dartmouth's own strict student AI citation policies. But he said, I develop ideas and arguments, conduct the research, evaluate the evidence and sources, determine the structure, prepare drafts, and decide what appears in the final text. I regard these technologies as assistive tools, he continued, adding they were no substitute for my research, intellectual judgment, or authorship.
SPEAKER_01Okay, so this is his public defense. Right. Okay.
SPEAKER_03And they um so here's the ironic twist pointed out by the student journalists. That came back as a hundred percent AI written. Ouch.
SPEAKER_01You would think like just your public defense, go train on your 2022 writing and and and use that or run it through a humanizer. The kids do it.
SPEAKER_03Right. I mean, uh I don't know.
SPEAKER_01All right. Now look, there's a reason we want to actually bring this in because we tend to stay away from this, like just rinse and repeat, rinse and repeat about academic integrity and what's real and what's not real. I was speaking with somebody, a school leader, um, let's let's say a department head in English, and he basically said, like, I thought my job was to help kids cultivate like deep reading and writing skills and master the art of writing, and I now realize that AI will just do that. And now my job is to help them think about critical understanding, you know, the piece that we're about to go in from Mollock. And so hold the thought on you're a very good question, I imagine, is like, well, what is his job? Right. But the fact that he was just willing to say aloud, like, all right, and we could have done this with with the provost. We just said, all right, look, that's not your job anymore. So fine, you used AI to do that. So this goes to Ethan Moloch's new piece, which is the overhang. One useful thing is his Substack, um, Mollock. And he says that while the public debate focuses on potential risks and the future of AI models, all this stuff in the news, current models like Astra and 5.1 Fable from Anthropic, you know, basically are creating flawless stuff. And he's not talking. About human writing. He is in part, he's basically saying it's making blender videos, it's making animations, it's making full models, it's building software, and it will write anything under the sun. And goes on to make the argument that we should stop getting into this kind of catfight about is it authentically mine? Like the AI will make the outputs. If that's the case, what is the job? What is the skill? What is it that we need as human beings? And so what is you know, I think if Moloch were to look at this, you know, provost thing, he would basically say, Yeah, it's clear. Like it will also go solve a you know a hundred-year millennium math problem.
SPEAKER_03Yeah.
SPEAKER_01So what what does he think the job is then?
SPEAKER_03Aaron Ross Powell, his response is they is basically because AI outputs are becoming fast and cheap, the human value is shifting away from the outputs towards four core differentiators: deep knowledge, wide knowledge, taste, and agency. So that is the provost fight. Like competing with AI on raw volume of output is just a completely losing proposition. Like human collaboration and guidance is the primary driver of value.
SPEAKER_01And that's what the English department chair is basically wrestling with. He's like, okay, this is like this chapter's done. Now what's the next chapter?
SPEAKER_03So basically it's having this deep knowledge, which means that you can detect errors instantly and then navigate the jagged frontier of these AI capabilities, right?
SPEAKER_01So Which is clearly critical, right? Because these stuff they make things up, they're making things up.
SPEAKER_03The algebra's incorrect, even at lower levels. And then when you get into the highest levels, like you've got to be an engineer or a physicist or whatever to be able to use these tools to create new discoveries. Okay. Wide knowledge is having knowledge across multiple disciplines to bring prompting together with precise domain technology. Um, you know, using the design terms, the filmmaking terms, even the knitting terms, right? Like, you know, you've got to have the right. If you don't know how to knit, you're not gonna use the words knit and pearl and slip slip knit and cable and all the different things, right?
SPEAKER_01I can connect it there to But to hang with the knitting for a second. It's like I might have deep knowledge in mathematics, but I have to have this like wide knowledge because I can use the descriptive language from the other fields, whether it's design or filmmaking or knitting, so the AI model is kind of like able to understand the way I engineer my prompts.
SPEAKER_03Exactly.
SPEAKER_01Okay, cool.
SPEAKER_03And then taste is just really important as this critical filter for selecting and refining and discarding the massive volume of the fast, cheap content um generated by AI to eliminate repetitive slap.
SPEAKER_01So maybe the taste question, maybe it's discernment, I don't know, is when that provost could have just made his defense. I mean, again, this is stupid because it's the whole argument is that none of this really matters, but it's like when didn't when to know what to use and not to use?
SPEAKER_03Aaron Powell Right. You know, like what what's cheap and tacky and what's like high fashion and expensive is is a matter of taste, right?
SPEAKER_04Like it.
SPEAKER_03And then agency involves actively testing the limits of AI capabilities as an explorer rather than waiting for instructions or official permission. So like being brave and testing things, but also saying, like, okay, I tried it now and it didn't work. Um I was just reading an article on AI ruining um knitting and crocheting because people are coming into these yarn shops and saying, like, here's the pattern I have and it's not working out. And these it's the beginner knitters, and all these knitting shop owners are like, that's because you bought an AI generated pattern. You didn't know it was an AI-gener, but you're now getting frustrated because you're trying to make this sweater with this pattern, and it's never going to work. And so yarn store store owners are seeing the impact of AI slop in their own way of like But but you could apply that.
SPEAKER_01So deep knowledge, wide knowledge, taste, and agency all would sort of make the difference as they build that skill set.
SPEAKER_03Right.
SPEAKER_01Who knew you could understand the whole world through knitting? It's amazing.
SPEAKER_03Aaron Powell Well, and that's where the real pattern makers are like, we're losing our business because people are going to cheaper patterns, but they're not actually getting what they paid for because they don't have these, you know, if you don't have all of these things, you will fall for AI slap, right?
SPEAKER_01Yeah. I'm gonna totally just change the intro to this is Chat you do a show about leading learning and knitting in the age of AI.
SPEAKER_03I'm all for it.
SPEAKER_01Okay. So, you know, look, we simplify this because you could easily look at the Dartmouth story and then look at Moloch's always incredibly thoughtful analysis and go, yep, makes sense. That chapter's over. This chapter's begun. Here's what we focus on. This is like nearly impossible to pull off because we have hundreds of years of valuing the thing, you know, that that we're gonna have to somehow make a shift around. I want to add one small addition, uh low for me to suggest something to Professor Mollock. Um, but I would add storytelling into the mix for sure. Like when I think about what Muse did for me this weekend, so in this podcast, I would argue that among the more important things that we do, besides our sparkling personalities and interactions, um, would be curation.
SPEAKER_03Like this is tons of AI stories. You pick the the right ones. Right.
SPEAKER_01So it's like, what do we talk about? How do we talk about it? So, you know, yes, Muse went in and did a bunch of like mechanical operational things for me this weekend that saved a couple hours of time. But research, curation, alignment, storytelling, taste discernment, wide knowledge, deep knowledge, all that kind of stuff were important. But I would argue that storytelling very much belongs in the mix.
SPEAKER_03Yeah. I mean, I think it's all important. And it gives us a purpose for school because that's what you have to develop the deep knowledge, the wide knowledge, the taste, and the agency. And you have to do that still through reading human-written things. So we do still need to read old books.
SPEAKER_01Yeah, we still need to write. I know. I I look, I do believe that writing is a way that we understand and we think and we process. Right. But the but we just we do have to have some start having some like hard, honest conversations about this stuff. Now finally, speaking of ways that we can develop deep and wide knowledge, um, we keep coming back to some more interesting developments that are happening with Google and other folks. But Google put out Google Research, which is from Google, and we'll be clear, uh, enabling teachers to create learning interactives with generative UI. This is a new study. Tell us about the study. And then you tested some of the little generative UI widgets or interactives or whatever they call them in their library.
SPEAKER_03Yeah. So they created this generative user interface experiment that empowers teachers to dynamically build custom interactive educational simulations for STEM subjects without needing coding skills. So it's grounded in the things we know about students like active learning principles, and it's built on Google's um Learn LM, which we covered a long time ago.
SPEAKER_01Aaron Powell Big custom model that's designed to support learning.
SPEAKER_03Yep. Um and so it incorporated strict pedagogical guardrails and agentic self-correcting loops to ensure simulations remain aligned with the specific learning objectives.
SPEAKER_01Aaron Powell Sounds like a lot. Stuff is good.
SPEAKER_03Yeah. So there is a public library of over 30 teacher-reviewed interactives across physics, chemistry, biology, math, and computer science. And there's some opportunities for pilot testing. But of course, um, my other assignment this weekend was to go and play with them. And I was playing with these last night and fell deep into the rabbit hole and was up way later than I had planned to be because um they were a lot of fun. So I, of course, went to factoring quadratic trinomials area models.
SPEAKER_01Of course, Cam, that's where I go to, right? Factoring quadratics always. Yeah, 100%.
SPEAKER_03I love a quadratic. You know, a good quadratic is always there.
SPEAKER_01I can understand why we're so close.
SPEAKER_03Anyway, um so basically when you open it, you go to the library and you open it, it tells you what the learning objectives are. So there are four um objectives, and then you launch this interactive. It can give you a tour, so you it tells you what to do, or you can dive right in. Highly recommend beginning with the tour because I had no idea. I'm not usually a tour person, and I was like, I'm just gonna figure it out. Had no idea what I was doing. Okay. So you should begin with a tour. But essentially, it walks you through five levels. First, you are like just doing you know, levels that are progressively getting you further and further down the road of understanding this thing. And you have these, you know, you're trying to factor the quadratic trinomials and you start with the coefficients and the constant terms, which are the easiest things to identify. There's slider buttons, there's tools on the right. So a calculator, there's formulas, there's theory.
SPEAKER_01Aaron Powell So would this be like, you know, because we've we've been talking about people vibe coding stuff in Gemini Canvas, is this similar or is it different?
SPEAKER_03It does feel pretty similar to something that's been vibecoded, except that there's more scaffolding in instruction.
SPEAKER_01Okay. So they build the learning science into the thing as opposed to someone just going into Gemini Canvas and saying, hey, make me a thing on quadratics.
SPEAKER_03Right. So I feel like the quadratic, if you just did that, it would probably be at level five and it's not scaffolding you through each step. Gotcha. Um so then I also wanted to try one that was much easier. So I stayed in the math lane because I'm a math person. Um and I went to decimal diner. Um so this is for elementary math, and it's about using decimal operations to calculate order totals and split bills, and you know. Um, so you are at a very basic, like the here you're not like, whoa, I feel like I'm in a diner. You feel like you're in an 80s video game.
SPEAKER_04Okay.
SPEAKER_03Um, and it shows like two items on a counter. So there's a taco and a soda, and it tells you tacos are 250, sodas or drinks are a dollar twenty-five. You have to calculate the total. So you just bring the slider over to the 375, and you're like, great. And it says that so that's the first level is just single item addition. Yeah. So you would like process that transaction. Great. So then it gives you one that's a little harder, multiple items. And so now you're gonna need to multiply and add. Again, that calculator is over there to the right. Then you start level three is you are splitting the per person cost. So now you've got to do some dividing. And level four gets harder, level five gets harder. So you're doing more complicated math as you go through, but you have all the tools there, and it is there's the ability to ask for a hint. Oh, yeah, and everything gets read to you too.
SPEAKER_01Oh, nice. I love this is multimodal.
SPEAKER_03Yeah, so it's multimodal.
SPEAKER_01With a good voice, by the way. Like in the Gemini Canvas voices, maybe they can be better. Some of them are not great.
SPEAKER_03Yeah. So, you know, it it lets you do all the things, it gives you the tour, you get you're walked through the different solutions. Now, is it the most exciting video game I've ever seen? No. Is it meant to be a video game? No, it's meant to be an interactive tool to walk kids through the steps. There are a bunch on science phenomena that I don't know enough about for me to evaluate. Um, the quadratic one was actually hard, like as I was going through it, but there are things as low grade as, you know, adding decimals and then moving into number line distance with integers and all the way up to quadratics and harder, you know, coordinating plane navigation. That's middle school, gravity and orbs and trigonomic functions with a Ferris wheel. So there are tons of different there's 30 of these. They are uh shareable and usable. So, you know, you could take this right from here, you launch that interactive, and you can share the link with your students. They can try it. Um, you're not, there's no chat bot or anything like that. You're not giving it any of your data. So they are kind of fun to play with. Um, just this morning I was on a Fulbright um international meeting talking about AI, and one of the teachers there was showing how to make HTML simulations for um classrooms he's from Cameroon, showing how to make simulations that don't require the internet. These do require the internet, but you could see how you could kind of take the same directions and do it in HTML.
SPEAKER_01Amazing. Thanks for doing two homework assignments over the weekend, but this is cool. I th I think you've you feel like you're this has been a value add for you.
SPEAKER_03Yeah, I think for teachers, yeah.
SPEAKER_01It's just another way, multimodal, another way to like visualize things and yeah, and as we're having all the debates about screen time, again, with any of these, I've been doing this with with people as they're doing some Gemini Canvas vibe coding of stuff. Like, think about SAMR. Like you could literally build SAMR into the prompt when you're doing when you're building some of the stuff yourself, right? That's the framework for thinking about whether or not the use of the technology is really just substituting something we could have done in analog, or it's meaningfully adding to the learning outcome. We'll put this into the show notes, including the library. They are opening up pilot participation for schools using Google Workspace for education. So uh these require teacher review and approval. But you know, something clearly like learn about. Like these are things that Google's really doing in the research lab, and I think we'll continue to pay close attention to them. Neat stuff. All right, Doc, we're gonna take a quick break, and then we're gonna come back from our with our two students who went to Boston, Baston, in order to, over the summer, um, with their superintendent to be part of that hundred student senate that we talked about earlier over the summer. Really psyched there actually, if you're listening to the podcast, they'll be on stage in the middle of the day today because they're gonna be part of our student leadership panel that we're doing where students are on AI squads or in leadership roles around AI in their districts, which we think is a good idea. So if you're enjoying Chat EDU, please uh like, follow, subscribe, write a review, all the things. So, Cam, you're doing all this stuff. I I just said in this podcast that you're gonna make short videos, you you've done it before, you're gonna pay the bills, um, you produce the like what can't you do?
SPEAKER_02Make toast.
SPEAKER_01True, true.
SPEAKER_02Hey, I'm Cameron Audia, executive producer of Chat EDU. As you likely know, Matt and Liz believe personal interest projects or pips are one of the best ways to help students bring agency and intellect to learning in the AI era. They also see pip as a safe, ethical, and powerful way for students to master AI. If you'd like to learn about Skills21's very own personal interest project platform and curriculum, check out skills21.org forward slash pip. And if you'd like to chat about pip in real time, email Liz at chatedu at edadvance.org. I mean, after all, she wrote a book about it.
SPEAKER_01So as promised, we're so excited to be uh joined with two students who went to Boston to work with a whole bunch of other students exploring AI and how to think about using it. These are two students from Stratford, Connecticut. We have Bina Mejia and Anthony Darbilly. Hey guys, thanks for joining us. Really appreciate it.
SPEAKER_06Thank you for having us.
SPEAKER_01Let's just jump in. Like, there's so many things to talk about with AI and students. Um, a couple of weeks ago, we covered a story from this thing called Common Sense Media. They do lots of research about students using AI. This summer they put out a big report like, How are students using AI? And I'm going to throw two statistics at you. They said about 60% of students use AI to do their homework for them, like answer this problem, uh, write this essay, do this thing, photomath this thing. And then 70%, which is a bigger number by my calculation, um, were actually using AI to help them learn something. So, like, don't give me the answer, but give me practice problems. Or here's the original essay that I wrote, here's my intro paragraph, um, help give me feedback. And look, sometimes those were the same kids, like 70% of the time, some kids were using it this way, 60% of the time they were using it that way. Like, I don't know, let's start with you, Bina. How does that jive with your understanding of like what you see with your friends? Those numbers seem like about right.
SPEAKER_06I would say so. I think a lot of kids see AI as a tool to save time on assignments that might just feel like they drag on. Um but I see a lot in my friends and myself, I use it to study or make practice tests, practice problems, whatever it is. Um, it's a very useful tool, but I think that a lot of times, like the 60% that use it for um just doing things for them, they start losing a lot of their critical thinking skills and they just start like relying on it a lot. Um, and it even shows up in school, you know. Um teachers don't know it, students know it. Um, it's more of a like, am I going to change this? Right? Am I going to try to do this better or in a way that's actually helping me for my future, or am I going to save myself a couple minutes? You know?
SPEAKER_01Okay. That makes sense. Anthony, what's your take?
SPEAKER_07I completely agree. I think when you use AI for those things, such as like having it do your work for you, or having it just do things that you can be doing, I think it creates like a dependence on it. You start depending on it for too much things. And like Vina said, you you lose that critical thinking skill that you have. And I think that just like it's something very valuable, that critical thinking skill, like it's something everyone in our generation should have. So losing it to like a robot that can do it for you, I think that could be a very big loss in our generation, just our community as a whole.
SPEAKER_01Yeah. Let me put one of you guys on the spot. Like if you had, I don't know if you use Chat GPT or Gemini, like what's the tool that you use? But if you were to look like, and I know school just got started, but if you looked at your chat history for the last couple days, can you give me like a concrete example of where you used AI in that 70% bucket, like to try to amplify your understanding or learn something?
SPEAKER_07So I think if you looked at my past release history, I've been I take AP Calculus BC this year, and one thing I've been working on a lot is working on some of my problems. And so I've been using it as kind of like a U tool, give me some practice problems and some study guides, because I have like tests and quizzes like every week. So I've been using it to give me practice problems, give it practice quizzes. Um, and then if I I give it my answers, tell it what I think is right. Sometimes it'll give me like, oh, this is wrong and tell you what I did wrong. But I definitely use it as like a tool to help me more than a tool to like just give me the answers and let me finish the homework.
SPEAKER_01But you know, do you ever worry? Because we know sometimes that AI is wrong. Do you ever worry in that kind of use case? You're like, wait, is this actually right?
SPEAKER_06Yeah, um, so what I've done before is I will do it myself, and if I get it wrong and it tells me it's wrong, if it doesn't explain it with some source or some kind of evidence for that, then I will ask it, like, are you sure? And then sometimes it has actually come back and it's like, oh, I was mistaken or something like that. Because again, like AI is gathering so many sources from everywhere. Um but no, I use it for more personalized education. So for me, if I say like I don't understand I don't understand a subject and my teacher is um not there or absent or it's just not during a school day, um I ask it to help me. And I think that's a really good way to use it because I'm not replacing my teacher's um teaching style. I'm just helping myself at home, like supplemental or learning.
SPEAKER_01Cool. What is your go-to tool? Notebook LM, Gemini, Chat GPT, what do you guys what do you use the most?
SPEAKER_06I use Gemini.
SPEAKER_01Yeah, me too. I use Gemini. Good choice, by the way, just because it's like, you know, tied to Google, you won't run out of tokens, all this stuff. All right, one more quick academic question. Yeah, so you're both kind of kind of like top of the list, you're doing quite well. There are students that are sometimes, you know, let's say you got a rank and you're in the top five percent, and there's a kid that's lower down, and they're using AI and they're getting results because the teachers aren't catching it. That's basically pushing them up. Is that bothering you? Like, do you feel this equity thing, or do you like you do you? What's your take on it? Yeah, you're both nodding. Give give me a quick, quick thought on that.
SPEAKER_07So from what I've seen, from a lot of things that I know, there are so many kids, just like not just in my school, but I see on social media, TikTok, Instagram, they all just use uh AI to get answers. And when they do that, obviously they get their straight A's and they get those good grades from using the AI. But I think seeing them like get better grades and better and higher ranks, higher GPAs, and those who actually work hard, I think kind of bothers me personally because I almost I feel bad. Like someone is over here working hard, uh, putting in hours of their day to study for a quiz and thinking like, oh, it's okay to make mistakes when there's someone else who can just quickly finish their assignment in 10 minutes and get a much better grade than them. And I think that just like upsets me for them and even for myself, because I'm sure there's kids who might be like better than me in a class or higher than me academically who might use Chat GBT or Gemini or any of those AI tools, and I just think it's unfair.
SPEAKER_01Yeah, this may be the toughest question, but like anything we can do about that, like Bina? I don't know.
SPEAKER_06Well, I know that for us, we uh our district is really good with our phones, especially now. We have a very strong phone policy, and even personal devices, like all of that is pretty monitored when it comes to test taking. Um so I think that a lot of the times like they'll be getting A's on all their homeworks, and then it comes test day, and the teachers taking phones and making sure they have it before they pass out the test, and we see that they do worse, and it's well, how do you understand the homework if you're not doing well in a test, right? So we do have reassessments and all of those things to keep you know our academics in place. Um, and as good as retakes are, our retake policy that we had, especially last year, we um like we saw less students like using AI for their tests because they were like, I can just retake it anyways, um, which I'm so glad that changed for us. Well, we can now like there's a process for it and it's not just just retake whatever you want.
SPEAKER_01Great. All right, well, this it's a good segue to Boston, right? Because at some level, like those kind of policy changes, like making policy changes inside your school on, you know, no phones, final assessments, got to show what you know. All right. So you guys went to Boston um with your superintendent, Dr. Borges, uh, who I have met and just super interesting leader. Um, how does all happen? Like, does she just reach out to you one day and say, hey, you want to go to Boston and talk about AI? Like, give me give give me the invite. How did it work?
SPEAKER_07Um, so Stratford got selected to represent Connecticut at at the conference because of how well our AI policies are. Um Stratford started developing uh policies like much earlier before other districts and they've grown to be very strong and they're very well so we were selected to represent Connecticut and Miss Borges reached out to my principal Dr. Graf and Vina's principal Mr. Delopiano and she told them to try and find a student who they think could best represent each school and Dr. Graf reached out to me and I think Mr. Delopano reached out to Vina and that's why we were selected. Very cool. And Bina, what did you think when you first got that invite?
SPEAKER_06At first I was really confused. I I won't lie um I was confused on what the AI festival would be and like is this you know like you know the culture festivals where they celebrate all these different cultures that's like how are we celebrating AI, you know? But I was really happy because I wasn't just representing you know my school I was representing Stratford and even outside of that like Connecticut like we were the two student representatives for our state which I think I was really I was really proud of you know I was like wow like like I was chosen for this great role and now I have to step up to it. So when it came to do like our research before we went there I was really on top of that to make sure like I was well educated before I got there.
SPEAKER_01Cool. So I guess two parts maybe we'll start with you Anthony and then we'll go to Bina. So what was the final product and then uh Anthony and then Bina how did you guys get there?
SPEAKER_07So the final product was one large bill split into four parts the first part was for students then we had teachers then parents and then administrators and I think there was there was a total of 100 kids two from each state and they had it divided into subcommittees each subcommittee was for one of those sections. I personally was on the administration subcommittee and Bina was on the student subcommittee. So in each committee we created each one created a list of bills which Bina will talk about more and the final product was all of these compromise and we all voted on all of them we combined them all into one big bill and then that was the final product.
SPEAKER_01Okay. And then Bina how long did this take like how long were you working at it?
SPEAKER_06Give us a flavor of like what did it look like to be on one of those committees well so we had a couple like a month or two I think to come up with these policies that we would submit on a Google form and then they surveyed all of our answers for what we would want to see in administration teachers parents students all of that and then wrote up their own like versions of that and then we got to decide do we like these or do we want to change something about them? What about them did we like did we not like and then we had to vote on that. So I was part of the student subdivision and it was I that was a really good and interesting one because again it impacts us the most directly right um so when it came to our discussions I got my amendment passed and it was about just when AI literacy would be introduced to classrooms and I really wanted to make sure that it was it was introduced when devices were introduced because um again we can't give kids computers and then give like expect them to know how to use them or how to um like understand what is false on it, you know? Another thing was digital literacy. It was just a big topic that we had discussions about but a lot of it came down to how well can you um fight for your argument or for your of your values in this and can you get people to understand or agree with you.
SPEAKER_01So it's interesting. So you arrived there's already like kind of a draft of a bill based on for each of these areas based on what you guys did as pre-work and then you're in there offering amendments and are like arguing or trying to build consensus. So interesting. Anthony did you also no judgment but did you also file an amendment I don't know if seems like a cool thing.
SPEAKER_07Yes once we got into our subcommittees firstly in the admission committee I think there was about like 15 to 20 of us and they divided us into like four groups of five and each group came up with their own amendment and then as a collective vote on one and sadly mine was not the one that was voted on. But I think the one that they got voted on I definitely agree with it. I think it was definitely an important amendment that was made. Well okay well here the the chat EDU audience is going to give you more airplay now so um what did get voted on that passed and what was your amendment that did not uh you know make uh so get across the line yes for the administration subcommittee the one that we voted on to pass was that administrators had to have like an equal balance when they use AI and that they need to when like accusing a student of mis of misusing AI or accusing them of using it for their homework or answers they need to have examples, sources and then allow the student to appeal what they did. So like they can say like no I did not do that and correct the administration. Yeah. And I 100% agree with that because I think that like with AI becoming such a big thing students are becoming like like accused like on accident so often and it's happened to me before so many times where a teacher will accuse me of using AI when I didn't just because they suspect someone is like very good writing. And so I think that having that appeal process is so important. And I think my original amendment in there was that administrators had to like have a certain list of how like when and how AI can be used for teachers. So like how teachers are able to implement it in their classrooms, how they're able to um let students use it. And while I do think mine is important I think the IPO process is definitely something that needs to be focused on first and then we can move on to the other processes.
SPEAKER_01Was whether or not teachers could use AI for grading part of your amendment? It was and we decided that teachers could not use it for grading.
SPEAKER_06Okay so some of some of the spirit of your amendment made it into the bill either way really interesting the teacher provisions actually focused a lot on that like what can teachers do with it right limiting it and making sure that teachers aren't using it to give feedback because that's something that we don't want to replace right the main one of our biggest talking points was never replacing human interactions or human qualities with AI just because it's easier or it's simpler. Because again being a teacher and the student teacher student dynamic is very important. Yep and it will continue to be important and an AI can never replicate that.
SPEAKER_01So it's so interesting that that you guys were kind of like and maybe those roles were helpful but you're really looking out for all these human relationships not just the student ones but the student and teacher ones. So you're really almost advocating on behalf of the teachers as well which is which is really neat. I'm curious in both of your areas like what's one part of the bill that came out that you don't necessarily agree with like I'm sure there was discussion and there were dissenting opinions uh if anything was there anything in each of your buckets or frankly on the parent or on the teacher side that you're not sure you kind of like still fully agree with even though you can celebrate the fact that it got consensus and passed.
SPEAKER_07So one thing that I personally disagreed with came from the parent section and what they were talking about is that parents need to also have a role in this and they need to like stop their kid from using AI when necessary. And I personally think that like when you have teenagers it's hard to really like know what they're doing at all times. You can't always track your phone and know what they're doing. Like I think it's very hard for the parents to have that role where they're going to stop their kid from using AI. Well I appreciate the effort of a parent trying I definitely think that that effort should be there. It's hard to have 100% full control over your child because if they're in school or they're in their room you don't know what they're doing. And so if they're using ChatGBT on like an on like a fake account or something like you really will never know. And I don't think that's the parent's fault.
SPEAKER_01It's funny I uh I'm just geeking out a little bit but I read the bill I also put the whole thing into notebook LM, listened to it as a podcast I can remember exactly where I was I was just really fascinated by this document. And so it's really fun hearing you guys talk about these pieces because so many of them were you know especially even being in that part about students have to take an AI literacy course before being able to use a tool I mean I got a ton of press and was all over adult boring LinkedIn social media. Sorry for all the LinkedIn people. Bina, what's something in your bucket in the student side or a different area where you're like I'm not really sure about that that passed.
SPEAKER_06I think it was again the parent section I had a really big issue with how are we enforcing that right um because nowadays parents are you know really involved in their own lives just as much as their kids but it's hard for them to you know supervise all of that. So I think when it came down to that argument I spoke on the Senate floor about how um Senate floor about how um it's going to be hard to implement that and that should be taught inside the classrooms where teachers can have an or overview of that because even though it's putting a lot of pressure on teachers it it also gives them more control over what's happening with their classwork if that makes sense.
SPEAKER_01Yeah. It does your parents both all your guys' parents would be very happy that you're like looking out for them. Yeah it was an interesting area right and it's nice that they were trying to speak to all these different roles can I throw two out to you that I wasn't really sure about it's funny I felt like an odd man out because all over you know again the press and the media everybody was basically saying and they're right that like adults haven't done a good job getting these kinds of policy documents and statements together and you guys came and you got it done which is amazing. But there are two that I had questioned about I'd love to throw them both out and have you guys you know feel free to debate me or tell me you know where you think the thinking might be right or wrong. So uh one on the student side there was a the language in there I think that said like students could use AI for like research and brainstorming and outlining, but you could not quote unquote write with AI. At least that's the way I understood it. And I wondered a little bit about that because I wonder like when you guys go off into the world of work, if you will in fact need to kind of collaborate with AI as a writing tool some teachers will have students write with AI and actually hand in the chat transcript where they're doing the writing with AI as like a skill. So there's one and you can either one of you can pick either one you want to look at. And the second, Bean and I were talking about this a little bit before the we started that you can't upload an assignment to AI. And I'm going back to that 70% thing and thinking, huh, like but don't does a student who's using AI to build their learning isn't it kind of good to upload that assignment so that the AI can help them troubleshoot and problem solve not necessarily be in the quote quote cheating bucket I don't know either one what do you think I think I saw you go like this. Yeah. So you you jump in. You take which one you want sorry Anthony.
SPEAKER_06So we had this conversation because it was a big debate actually about well what if they source like they they cite their sources of it um but then it wouldn't be their source it would be AI's source. So it was a big discussion but we ultimately agreed that it just like writing with AI isn't helping you learn right it's you can't ask AI to do simple math for you. We made this comparison to when you're learning addition you can't just jump into multiplication or algebra or trigonometry because you have to know the basics right so if you are okay with writing the basics being able to write an essay on your own and you want input from AI that is what we saw as okay but using AI to just do it for you and turning your brain off for an assignment that you just don't want to spend time in that isn't helping you in any way right so again it's the what's helping what's hurting you um yeah you have to know the basics so can you do this yourself or are you just using AI to do it for you without learning how to do it.
SPEAKER_01How hot was that debate like was it it was hot it was hot. How long did it go on for?
SPEAKER_06Pretty long but I think the longest one was about equity and like what about schools that don't have the resources and you know talking about all of that but all right cool we'll come back around to that.
SPEAKER_01So Anthony being a pick so now you inherited the um can students upload an assignment to AI? So any any thoughts on that?
SPEAKER_07I think the act of submitting an assignment into AI is actually very controversial controversial and it has a lot of parts to it. Yeah um I I've seen kids like say right next to me will go download an assignment as a PDF and put it directly into ChatGPT, Gemini Cloud, any of those and I see that and they get the answers immediately sometimes even mix a Google Doc with the answers they can just submit. And I'm just like personally I I look at that and I'm like you learn nothing like you don't know anything that happened. And so that's why I think if you are looking at an assignment you have a question about it like let's say you're looking um at a math assignment like I do and I'm working through the problem and then halfway through I get stuck let's say factoring a problem. I think it's okay to ask AI, hi I'm stuck on factoring this what suggestions do you have for me that I can do so I think you're inputting part of an assignment or part of a problem into it to get some feedback and some tips. I think that's okay. But I think to take an entire assignment that was made by your teacher who put a lot of work into that assignment to just go and directly put it into an AI and then have a Google Doc ready to be submitted be generated for you literally like in a minute I think that's just it not only doesn't like harm the student, but I think it's also just disrespectful to the teacher because there's someone who put so much work into assignment for you. They're doing it teach you so that you can understand and you're just completely disregarding it.
SPEAKER_06And if I can add one quick thing is um I think uploading an assignment is very different than uploading your notes from a classroom or a lecture because the assignment is something that you were tasked to do and then the notes can help you uploading your notes and then having it either read it to you or re like teach it to you it's super helpful. I think some people just learn differently and personalizing your education shouldn't be seen as evil or bad um because you're doing it to help you. You're not doing it to um just get out of doing your work.
SPEAKER_01Cool. Yeah I mean I was wondering about the upload because I was like oh maybe it's part of the that 70% to like learn how to study or grow but I think you guys both just convinced me against my own thoughts on both of those. So well done I I wouldn't have I wouldn't have lasted on the Senate floor. All right listen before we wrap like I'm sure this was super interesting you're also hanging out with 98 other teenagers from all over the country like that had to be interesting. I'm curious was it interesting like what were your big takeaways from like hanging out with these kids from everywhere from Hawaii to New Jersey to Florida to Wyoming you know who'd you hang with what was interesting?
SPEAKER_07Personally for me I think most interesting was how different everybody was like I learned things about other states and like people that I had no idea even existed. So Stratford with our such like a well-written AI policy and everyone realized that some states like don't even say a word about AI they don't have any policies regarding it at all and that shocks me because I'm like like my our AI policy helps me students and other teachers so much so the fact that some had none was honestly just shocking to me. So when I'm talking with these fellow senators from all around the country I'm just like wow like it's just so surprising to me like one of my fellow senators was saying how they how in their state that they just completely disregard AI. There's no policies about it. Obviously a teacher will like check if you've done it but there's no specific policies like that are made by a district board or anything. And I think that shocks me like a lot because I think our policy has shaped so much like structure in our schools but I think it's definitely helped a lot.
SPEAKER_06Cool Dina how about you I'm just thinking of the friendships we made honestly um it was so interesting all of them were so different but we somehow ended up blending in so well um I think like New Jersey Arizona they had amazing people and I'm so glad they sent them over um but again it was just making these connections and having these conversations that we had uh we would stay up very late in the dorms just talking um they had communical communal spaces so we were just bonding and we played some games but overall we had really meaningful conversations even outside of just AI right it was coming together for something like this was so awesome and I wish everyone had the opportunity to do that and I wish you were there for longer but you know I'm grateful for the three days that we did have have you guys stayed in touch with some of the other students yeah we actually had like a group chat with some of our like listeners and we still have like texts and talking it to this day.
SPEAKER_01That's cool. Hey listen so last thing I think um you're in a district that's got like this great policy it's really interesting to hear you say like just clearly from a student's perspective that having that policy is important because frankly you know people listen to this podcast around the country a little bit around the world there are lots of places that don't so I think it's just really really helpful to hear you say that it's helpful to you it's helpful to the teachers. So you had a good one that said you went and hung out with all these other students you did all this deliberate thinking did you or Dr. Borges are you guys together come back and like apply or change or modify or propose making any changes in Stratford's policy based on your conversations in Boston?
SPEAKER_06So me and Anthony are part of our AI committee that we are we're trying to basically come up with a curriculum that could apply and help us achieve this AI literacy that we talked so much about in Boston. So this year we're gonna work on finding the resources and finding what we do want to see in our schools and the changes that we want to make. But we have spoken to representatives from the states we talked to representative leaper um yeah so we've been reaching out a lot just to see what we can do how we can actually make change after passing the bill.
SPEAKER_07Amazing yeah me and Bina are also working with Dr. Michelle Eckler and we have plans for this year because we're both like you said on a digital wellness task force for AI and that task force is partnered with um Pinterest and ISTE. So we're gonna be there's like talks of like we're gonna try and write like a curriculum for a new like AI literacy class. So me and Bina and Dr. Michelle Eckler and that whole committee are going to try and collaborate together to write this curriculum and hopefully get it approved to hopefully start in the next few years.
SPEAKER_01Amazing well definitely keep in touch with us and let us know how that comes along and again I just thank you both for representing Connecticut so well for representing your generation frankly so well and I really just appreciate you taking some time to be with us and our audience and we're just incredibly grateful for you. So thank you. Thank you for having us.
SPEAKER_06Yes thank you so much this is great.
SPEAKER_01Awesome Liz I was bummed you weren't with me for that one.
SPEAKER_03I'm so mad I missed it.
SPEAKER_01But I'm excited But you're you're facilitating right now you're facilitating the panel with the students I'm very excited to facilitate the panel.
SPEAKER_03I'm so glad we got to meet them and have them at our conference but I'm just really bummed that I missed the interview.
SPEAKER_01Well you were doing important work on the Ask in Houston so it's true. All right so uh Liz, time for a bright bite a place where AI is making a meaningful contribution to health society, culture, the environment and the like and this is a great example from the Gates Foundation.
SPEAKER_03So the Gates Foundation's 2026 Goalkeepers report focuses on the transformative potential of artificial intelligence to act as a global equalizer or significantly widen existing health and educational divides. So they have um backed this effort with a billion dollar commitment from the foundation to because if we just leave AI purely to market forces, it's going to result in advanced tools serving the wealthy population first and then leave developing regions behind. And so to prevent that the report outlines three things. One is to expand AI language capabilities to cover lower resource languages so more less represented in the models and and just smaller percentages exactly right to ground tools in local data and realities. We can't just keep having data coming from large industrialized nations and investing in human infrastructure and access for frontline workers like teachers, doctors and agricultural advisors around the world again, not just in the wealthy democratized places where it's already I mean he says the question isn't whether AI will transform the world it's who will benefit from the transformation which I just love. You know we, you know I talk about equity all the time and so um and just being on this morning with colleagues from Cameroon and India and seeing the differences in access. So you know the the man from Cameroon Najee Patrick was a Fulbright to the United States. So our my colleague on the Fulbright Nanji Patrick was a Fulbrighter from Cameroon to the United States where he was here and he's like I felt like I was in a space rocket like everything is digitized. Everything you touch in American classroom has, you know, buttons and you know screens and everything and then I go back to Cameroon and he showed a picture of his classroom there, which is literally dirt floor, no tech, you know, and his presentation was like how do you still teach about AI when there is low technology available or AI of you know little AI available and he gave great examples of how to use um you know when the teacher can still have access how to use that and then when kids go home and have access even if they don't have it in school how to do things like that to expose kids to AI. And that's exactly the kind of communities I'm thinking about is like just seeing non G students in Cameroon was really cool.
SPEAKER_01Yeah benefiting from this Gates investment for sure.
SPEAKER_03Yeah.
SPEAKER_01Neat okay well doc we'll see you at the conference it must in fact now be time to roll roll roll the credits. Wow three rolls see you next week.
SPEAKER_03Bye bye got thoughts on this episode we want to hear then email us at chatedu at edadvance.org that's ednc eorg and hit subscribe rate and review wherever you listen. Chat EDU is proudly produced by EdAvance Connecticut's regional education service center located in Lichfield Connecticut huge thanks to the crew Andre Angelo who edits and produces the show and helps us sound great Cameron Audia our executive producer who makes us look great on camera and is our resident rock star and Gary Robinson whose time and talent keep things running with all things graphics, thumbnails and more. And a special nod to Jonathan Costa, our executive director and the reason this podcast exists in the first place. Most of all thank you for listening and watching we don't take it for granted.
SPEAKER_00Hello everyone my name is Jonathan Costa and I'm the executive director at EdAdvance. I want you to know how much we appreciate our listeners. If you enjoy the podcast feel free to explore EdAdvance's other AI services. These include EdAdvance AI Studio an AI powered platform that helps educators, related service providers and leaders develop high quality compliance student-centered IEPs, 504 plans and MTSS plans along with progress monitoring, reporting and much more. We also offer customized professional learning of all kinds conferences, our middle and high school student AI literacy course and our educator and leader microcredential co sponsored by the Southern Connecticut State University Office of Workforce and Lifelong Learning. Just visit skills21.orgslashai to learn more. That's all for now and thank you for listening to Chat EDU