00:02.37

Eric Stano

Hi, this is Eric Stano with Magic EdTech, and this is the Tech in EdTech podcast. I'm joined today by Chuck Melley, the Senior VP of Policy Advocacy and Communities with CodeAI, and we'll be talking about AI literacy and changes in technology in the classroom. Welcome, Chuck. I appreciate you joining me today.

00:23.19

Chuck Melley

It's great to have you here, Eric. Thanks so much.

00:25.39

Eric Stano

Yeah. Well, thank you. And Chuck, for our listeners who may not know you or haven't encountered you on a podcast in the past, could you tell us a little bit about your journey

into education policy broadly speaking and the work that you now do at CodeAI.

00:43.38

Chuck Melley

Oh, I sure can. Yeah, definitely, Eric. It was a bit of a winding road, but with a singular theme. You know, growing up, my family was all about public service. My parents were teachers. Many of my uncles were in the military and, you know, some police officers and some public officials as well. So, you know, I've decided, at least from a very young age, I decided to, you know, dedicate myself to public service. But, you know, growing up, I realized I don't have enough courage, or the ability, or talent, to be a teacher, and my interest, kind of, lies in politics, you know. More so than the actual teaching, being in the classroom. So, you know, I focused on that. I learned that, I think, when I was probably around 17 or 18 years old. And so, I decided to just jump in that way, you know. I went to college at night, worked on the Hill during the day, worked, you know, a bunch of house campaigns. And then after a while, I worked for- I was actually a presidential appointee working at the Department of Commerce for two terms of administration. And it was really funny when I was at Commerce; I guess it was 1999, 2000, or something like that. That's when the internet became the big thing, right? And so, I was…

01:51.78

Eric Stano

Right.

01:53.10

Chuck Melley

… kind of one of the guys focusing on that and, you know, believe it or not, I actually worked on trying to figure out how we can close the digital divide. Something, what, 27 years later, we're still trying to work.

02:04.91

Eric Stano

Still using the same words, in fact.

02:09.10

Chuck Melley

Exactly. Exactly. It's really funny. And well, then when the administration ended, I thought to myself, well, you know what, you know, politics is fun, but it's changing a lot. And I was really interested in education overall, and so I focused on that. I went to work for ETS, Educational Testing Service, and I was there for a long time, you know, doing what I do, you know, government relations and media relations and things like that. And then, I guess it was probably around 2000, I don't know, 13. That's when I realized that, you know, education was turning. It was actually at an inflection point, right? It was going digital. And I'm sure you remember that, you know, it's… everything was, was when I first moved to Pearson.

02:53.48

Eric Stano

Well I appreciate you, uh, revealing my age in the course of that, but yes, I do remember it.

02:59.05

Chuck Melley

Yes. And it's interesting because at the time, you know, I'm thinking that, well, now at this inflection point, sure, I can continue to work at ETS, which is a wonderful organization. It still is, by the way.

03:10.50

Eric Stano

Of course

03:11.11

Chuck Melley

But I thought I could work… If I talk to Pearson and go over to Pearson, they're a much larger organization and have… I could potentially have a larger impact, right? And I could be there during this digital transition, you know. From moving from, you know, forklifts and warehouses to bits and bytes, right? And so, that was great fun at Pearson for a number of years, you know. Running global government relations and working with a lot of really, really terrific people. And then, you know, a few years later I realized... well, I think we all saw this coming. AI is going to change education overall, and we're really going to be…  we're at another inflection point, you know. And so, then I started talking to the CodeAI people, and they were kind enough to, you know, let me on to help out; help navigate them through this transition period and into this next, I guess, in the next iteration of where we're at.

04:15.42

Eric Stano

Sure.

04:16.09

Chuck Melley

Where we're going to be. And so.

04:18.48

Eric Stano

Sure.

04:19.40

Chuck Melley

In Code, I don't know. I don't know if you all know much about Code, but maybe I could just do it real quick.

04:23.61

Eric Stano

Of course. Please tell us a little bit.

04:25.56

Chuck Melley

Yeah, it's a nonprofit organization. It used to be Code.org. Now it's CodeAI. And its main mission is digital fluency for every student, you know, regardless of zip code, country, you know, wherever. At CodeAI, we build the, you know, the curriculum and train the teachers and advocate for policies to make that possible. I mean, one of the reasons why I'm so glad to, you know, let me join is the impact, right? I think…

04:51.01

Eric Stano

Right. right

04:51.96

Chuck Melley

… like over 150 million students reached two billion plus hours of learning, and now we have the…used to be the Hour of Code. Now it's the Hour of AI. And, you know, in every state and in the country and in the 900 countries around the world. I mean, it's super proud, super lucky, super fortunate to be working with, you know, partners, departments of education, ministries of education, you know, to try to figure out how we can expand digital fluency, you know, among the learning population these days and moving forward.

05:25.10

Eric Stano

Yeah, that's great, and actually, just for full disclosure for any of our listeners, Chuck and I share a couple of intersections. We were both, in our careers, we were both at Pearson. I was in politics, but in an academic sense at the time as well. I was the political science editor, so I'm interested in what Chuck does in that way. And also, working at Magic, I share some of the mission that he just articulated for CodeAI. Magic's mantra is to provide digital learning for everyone. And so, we're also interested in equity in that way and many of the things that you were saying. So, really appreciate that walk down memory lane, Chuck. Now, focusing on the work that you're doing now, you know, you've always focused, or often focused, really, on helping more students, you know, understand and, you know, bridge the gap in technology. You want them to understand what it is they're using, not simply consume it. What brought AI literacy to the center? I think we could probably intuit some of the answer, but I'd love to hear you just talk through how AI literacy really ended up at the centerpiece of what you do.

06:44.60

Chuck Melley

Well, yeah, I mean, especially over the… and first off, I just want to note that, you know, when I was at ETS, and when we were working together at Pearson, we used AI. But it wasn't large language models, you know. It's a different type of AI.

06:56.31

Eric Stano

Right.

06:57.11

Chuck Melley

Right? We used papers and things like that.

06:57.44

Eric Stano

That's right. That's right

07:00.11

Chuck Melley

And so, relatively recently, that this new iteration of AI is really what we're talking about now. And this kind of AI - the large language models and other uses - is much more impactful and used, you know, by everybody. Just think about it. Students are using AI everywhere, all day, every day, you know. Well, I guess in many places after, between the bills, not during the bills, right now, in a lot of places.

07:23.74

Eric Stano

That's right

07:28.84

Chuck Melley

And the thing is, it's, you know, teaching in most places hasn't really caught up to them yet, right? You know, there's a study I read; I think it was 84 % of US students use AI tools, but only 16% of high school leaders say that students learn about AI in school.

07:49.09

Eric Stano

That's right.

07:49.05

Chuck Melley

I mean, you know, kids are using it, but they don't know about it. So, we think it's super important that they really learn about it. And then, especially for all kids, because right now if it's only the, you know, the handful of students, the 60% that are learning about it, it's probably in the most well-resourced districts, you know. And so…

08:09.87

Eric Stano

Absolutely.

08:10.64

Chuck Melley

… yeah, exactly. So, under-resourced schools, or the ones with no computer science teachers on staff, you know, those kids risk being left behind.

08:20.63

Eric Stano

Well, and on that, you know, certainly there is the disparity of those who learn about AI and develop AI literacy. I'd be curious, though, if you could talk about what AI literacy actually means beyond knowing how to operate, say, a chatbot. What do we mean when we talk about AI literacy among students?

08:44.22

Chuck Melley

Oh yeah, definitely. I mean, using a chatbot is the easy part, right? Putting in a prompt. That's it. But the skill that really matters is judgment. It's, I mean, it's knowing what that LLM is doing, knowing when to trust it and knowing when to question it. And the way you do that is by, you know, building in a deep, but actually a good understanding of how it's built and how they work. I mean, students really need to be able to understand how to direct it, question it, create with it, and then, maybe, help shape it moving forward.

09:21.61

Eric Stano

Right. Does that suggest that students, you know, need to understand the algorithms that they're engaged with? Or does that mean that students need to be… to develop skepticism and critical thinking skills? It sounds like there's, maybe, a sort of a tangle of different things that students should really be developing. And/or helped to develop in this context. Can you say a little bit more about that?

09:50.77

Chuck Melley

Oh yeah, sure, definitely. I mean, not every student has to know how to build a model, right? 

09:56.87

Eric Stano

Right

09:57.44

Chuck Melley

That's down the road. But now they really have to understand. I think it's important for them to really fully understand how they're built and how they're made. The LLMs provide responses that are based on data that is input by humans, right?

10:15.58

Eric Stano

Right.

10:15.65

Chuck Melley

And fill it. It reflects all the accuracies and inaccuracies.

10:22.53

Eric Stano

The biases, so yeah.

10:24.88

Chuck Melley

And one way to think about this, as far as what they should know and what they shouldn't know, is the OECD-EU AI Literacy Framework. I know that's a mouthful.

10:36.47

Eric Stano

That was a lot of letters of the alphabet, certainly.

10:39.65

Chuck Melley

Right, exactly. But what the OECD did… and this is actually CodeAI worked with them on this before I joined. What they did is, you know, with all the different countries and all the different experts in CodeAI, they put together a framework of what students should know in a durable framework. And, you know, to really help equip them to not only live but survive, but also thrive. I mean, it's really interesting because, I mean, there's like knowledge and skills, right? And while things, you know, are changing and technology is changing all the time, there are certain things that will stay the same.

11:10.95

Eric Stano

Mhm.

11:17.07

Chuck Melley

And to your point, it's about, you know, just understanding how AI systems work. Like, you say the algorithms and things like that. But then, also, the societal and ethical implications, you know. I mean of AI and what potentially it could do. But then also on the skills side, I mean, it's really important that they've developed the fundamentally human capabilities, such as computational thinking, creativity, critical thinking. There's... those things that make us human. I think this will make us more able to, and we develop them further, will make us each individual more able to really work well with AI and move forward because it's, you know, especially as AI advances, it's in changes. So, I think it's learning the skills; what it's about. The skills: how to direct, interrogate, and, I guess, audit with the AI effectively. So that whatever tools come down the pike moving forward, you know, they are able to work with it, manage it, and use it to accomplish their goals.

12:27.43

Eric Stano

And you said a lot there that's… I really want to dig into, you know. You mentioned frameworks and, you know, also a variety of skills and concepts. And you used the word durable as well. Relative to the frameworks and the skills and the concepts, could you, at a high level, again, if there are parents or elementary school teachers or, you know, teachers in middle school or high school. What sort of concepts and skills, if you could, yet again, at a high level, what should we expect of an elementary school student versus a high school student? How might they have or develop AI literacy that's appropriate for their age group? What should teachers be focused on helping them do relative to their engagement with AI? And then I'll get to the durable piece that I wanted to follow up on in a minute.

13:26.10

Chuck Melley

Oh, great, great. Yeah, that's a really important question because there's a huge distinction between, you know, like, say grades three through five, six through eight, and in high school, you know. I mean, because three through five, it's just introducing, you know, the students to the AI and machine learning to demystify the technology. It’s not magic, right? It's pattern-based logic….

13:45.56

Eric Stano

Right. Right.

13:49.46

Chuck Melley

…you know. Rather than being either something just magical or random, quite frankly, it's just the concept of understanding the patterns. And you know what? This doesn't have to be taught on the screen, not all of it. It's really about thinking and learning, and to a certain extent, math, if you want to think about it from a process standpoint.

14:09.91

Eric Stano

uh-huh

14:10.84

Chuck Melley

You know?

14:11.57

Eric Stano

Yeah.

14:11.72

Chuck Melley

And then, I mean, in the middle grades, it's really starting to put, you know, digital fluency, because that's where we're going with this as far as digital fluency.

14:11.74

Eric Stano

Yeah

14:20.87

Chuck Melley

And starting to put that at the center of, you know, of the instruction and the learning process. So they, you know, know not only how to use AI, but also how to think about it too, you know. In what context should they put it, really, you know, as far as going back to critically thinking about the human piece. And then, once you get to high school, I mean it's really going from understanding how AI works to actually being a creator, right? To building apps with it to using it to build new and different things. And then, also, you know, start to really look under the hood as far as how it works. And so, maybe, lay the groundwork, as I mentioned earlier, and become a shaper.

15:07.82

Eric Stano

It's interesting. So actually, I'm probably characterizing this very reductively, but in elementary school, it sounds like, you know, sort of real recognition of what it is and what it is not. In middle school, you're beginning to apply some critical thought  to its usage, and in high school, it gets a little bit more constructivist in nature, where you're actually employing and applying it, knowing what it is and having thought critically about it. That's where you're able to actually generate work using it and interfacing with it. Is that a fair objective summary?

15:46.46

Chuck Melley

Yeah. Yeah, that is it. That is… that it's great. That is very accurate. 

15:51.20

Eric Stano

Great. Well, then let me return back to the durable comment, because I'm really interested in that. Because, obviously, you know, you worked at Pearson, as did I. The development of content and curriculum, you know, has not traditionally been done at that lightning speed. And certainly technology, you know, grows and develops at an exponential rate. What, in that kind of context and in an educational setting, what sort of concepts and capabilities are durable enough to teach, you know, amid a technology landscape that evolves so quickly? What do you tend to focus on?

16:36.46

Chuck Melley

Yeah, I mean, as far as fundamentally, you're 100 % correct, you know. I mean, the tools we use today are gonna look so different in five years, two years, who knows, two months, maybe. I don't know, you know. But, I mean, how AI works actually, you know, it won't change. It learns. Like, I kind of touched on before, it learns from data that's made by humans.

16:52.96

Eric Stano

Yeah.

16:54.05

Chuck Melley

And, you know, it makes predictions, and it can be wrong or biased. And you have to check that it works to make sure that it happens, you know. That it's accurate. And, as far as the, you know, the durability piece, it's the habits that we want to develop: durable habits, verifying outputs, asking where information comes from, deciding when AI is the right tool or if it's not. And, you know, and yeah, going back to the OECD-EU AI literacy framework, they have, you know, they did a really good job as far as domains, you know. This is from our work, and, you know, pure narrative, you know, the, of course, frameworks have domains, you know, of learning and, you know, one is engaging, one is creating, one is managing and designing.

17:36.16

Eric Stano

Sure.

17:40.70

Chuck Melley

And underneath that, you know, there are different skills in each one of those, you know, and they're aligned along the journey so that, you know, as all of these different models change progress, those who are, you know, digitally literate, they have the ability to function and succeed in whatever environment comes their way. It's thinking. It's really a way of thinking and approaching things.

18:14.17

Eric Stano

No, it's a soapbox that I find myself perched atop a great deal of the time. Having come out of humanities education and having published largely around the social sciences and humanities, I find myself agitating for the development of critical thinking and skepticism and just those soft skills that I think are the ones you learn with the humanities and social science education, not exclusively, but certainly there, and, you know, I've promoted the value of those skills, particularly in an AI-infused world. I think it'll be things like judgment and critical thinking that are going to be the skills that carry the day in the future. Let's shift topics to one that, again, both with our political backgrounds, both you and I, I think, enjoy, which is policy.

19:08.06

Chuck Melley

Mm hmm.

19:08.84

Eric Stano

I'd love to hear your reflections on, you know, if a state or district says it wants students to be prepared for an AI-shaped economy, what are the policy levers that are available to them to make that real? Things like learning standards, graduation requirements, you know, policies at the, you know, the school level or state level, and things around course access, teacher preparation, funding. But again, there's, I'm sure, a whole tangle of things that you know better than I from having been in the real policy world. What's available to a state or a district to, you know, prepare their students for an AI-shaped economy.

19:54.87

Chuck Melley

Yeah, I mean, you know, all those that you mentioned and more, you know, all of the above.

19:58.93

Eric Stano

Yeah, sure

19:59.58

Chuck Melley

And there's no single lever that works alone. I mean, you know, for example, from what I understand, before I joined CodeAI, you know, they worked really hard to get computer science policies adopted in the U.S. in those hundred and 90 countries around the world, and they say that over 350 policies worked because they were stacked. There were standards, dedicated funding, teacher certification, professional development, course access, you know, in some cases, graduation requirements. And so, one of the things that we're doing, and what we've done and continue to do at CodeAI, is to help the, you know, states, countries, and, you know, municipal bodies with this. So, with this process, we developed the ten policies for digital fluency. And, in there, you know, it's on the website, things like, as simple as create a statewide plan and guidance for digital sciences education, you know, establish standards and, you know, establish dedicated digital science positions in state agencies, and then reporting accountability, and then, of course, you know, assessment, you know, you're going to see how kids are doing if they're learning.

21:15.01

Eric Stano

Certainly.

21:15.82

Chuck Melley

We think they should learn. And, you know, offer it to all the schools and all the students and, you know, allocate funding for it. Cause I mean, well, a big thing is, I should say before allocating funding, is professional development. Because I think that is key. Going back to that thing that I said, where just a very small portion of students of schools are actually at, you know, teaching this professional development, is just crucially important. And so, for that and everything I've mentioned before, there has to be the funding and, you know, the determination to get it done and do it right.

21:54.91

Eric Stano

Right. And actually, I think you've probably answered this a little bit in your comments about the stacking with policies there, you know. There's a risk of writing policies that are so broad; they change nothing, or are so tool- or application-specific that they're, you know, obsolete almost immediately. Would it be through that, sort of, staffing of policies that you were referencing, that would enable people to avoid those kinds of extremes?

22:27.76

Chuck Melley

It would, it would, you know. And we have the guidelines here, but, you know, education is local, right? What I would suggest is that, you know, the policies that they adopt to avoid those two extremes you mentioned are to just focus on the outcomes.

22:33.36

Eric Stano

Right, right, of course.

22:45.35

Chuck Melley

Focus on the, you know, the concepts. Certainly don't talk about the focus on the tools, you know. Because we touched on that, tools will change, but, you know, outcomes and concepts, and then…

22:45.56

Eric Stano

Yeah.

22:51.28

Eric Stano

Right.

22:57.93

Chuck Melley

… I said this a second ago, but pair every aspiration with a resource, you know, a mandate without…

23:04.60

Eric Stano

Nicely said.

23:05.58

Chuck Melley

…one thing is there's nothing, right? 

23:09.32

Eric Stano

Right. I support an unfunded mandate in our other parlance.

23:14.74

Chuck Melley

Right. Exactly. Exactly. And then, it might not get done. And then, the key thing is also to build a review cycle, you know. And, as they set up their policies, put them in place and, you know, they can update them, you know, and just to see how they're doing so they can update the guidance, and update their processes without reopening statute, you know, without changing laws or standards or something like that. So, I think it is…. it's, yes, definitely outcomes, concepts, funding in, you know, be flexible moving forward.

23:46.53

Eric Stano

Right, right. In this context, you know, presuming there's funding focus on outcomes, everything you've said, and you referenced this earlier, but I'd be curious if we could go a little bit deeper. What does realistic teacher preparation look like? You referred to professional development, you know, and what does it look like, and what should school leaders stop expecting teachers to figure out on their own?

24:17.59

Chuck Melley

Oh, yeah, definitely. I mean, the teachers are key. As I mentioned, my parents were teachers; my wife's a teacher.

24:20.59

Eric Stano

Yeah

24:23.97

Chuck Melley

And so, it's proven that the teacher-student relationship is the most important thing in the education process, right? And so, we have to, you know, understand and let folks know that teachers do not need to be AI engineers to teach this, right? They need conceptual grounding and support, and that support being ready-to-use curriculum and instructional materials. And so, you know, so that they're not building lessons and lessons from scratch. And they need, you know, structured professional development, because we can't, well, we have to stop expecting them, quite frankly, to self-teach on nights and weekends. And, you know, buy two tools on their own and vet them on their own, and write their own AI lessons. So, I think it's giving them the guidance, the support, and the grounding they need to succeed. Because, you know, pedagogy is the most important thing, and teachers are masters at that, and those who are gifted should continue to do so, and we should continue to support them. Oh, and I just want to say one other thing that is, you know, at CodeAI, our teacher training, you know, included AI content for years, but it's specifically designed for teachers without a computer science background. So it's great if they major in computer science and the classroom. It's great if they worked at a tech company or something like that, but they don't have to be.

25:49.69

Eric Stano

Right.

25:53.43

Chuck Melley

Because back to the point about pedagogy, right? 

25:56.71

Eric Stano

Right, right, right.

25:58.17

Chuck Melley

And let teachers teach. Give them the tools and let them teach, right?

26:01.48

Eric Stano

Yeah, and given the success of any of these policies or any of these implementations, you know, is sort of naturally seen through, you mentioned outcomes, through the outcomes, the performance of the students, etc., through sort of review cycles on how those policies are working. But, you know, as I listen to that and I think about it, you know, in some cases, in many cases, you know, if it's not working, well, you know, we don't really know until, you know, a whole lot of students have moved through that content, you know. That suffer the consequences, so to speak, of that. What's an early sign that you think a district could look to, to see if they produce an AI plan? Or an acceptable use policy, but has not built the capacity to implement it well. Or has not thought it through in a way where it could be executed well. Are there early signs so we can get in front of it better and, you know, preserve the potential outcomes of the students?

27:19.92

Chuck Melley

Oh, 100%. 100%. And it starts by looking at the plan, right?

27:23.47

Eric Stano

Yeah.

27:23.49

Chuck Melley

I mean, if the plan is all rules and, you know, no instruction, it just says, you know, like it says what students can't do with AI and, you know, not where they learn about it, or how to create with it. That's a big red flag, right? You know, it…

27:40.90

Eric Stano

Right.

27:41.49

Chuck Melley

…it's really about… oh, and then, of course, there's no budget line for professional development funding, a dedicated person or, you know, a function in the department of education or the SEL, or that owns AI literacy; that's a big flag. So, it's getting under the hood of that plan and seeing not only if… carrying on this hood analogy, if the engine is big enough, you know, the cylinders, but also if it has gas, you know, that it can run.

28:15.07

Eric Stano

It's… It's funny you're using the automotive analogy, or speaking to a New Yorker who's never owned a car. So I'm presuming it's apt. And I'm sure all of the listeners own cars, so it's probably really resonant with them. So I'll presume that to be true.

28:32.27

Chuck Melley

Maybe.

28:32.23

Eric Stano

See?

28:33.02

Chuck Melley

Maybe not with electric cars, perhaps, you know?

28:37.57

Eric Stano

No, and actually what I like about what you were saying is it sounds like… and this, maybe, is something that is common knowledge in the policymaking world, but it sounds like you were talking about us really having a rubric framework. Or, more simply, it's sort of a lens through which to look at and evaluate any plan to see if it holds water. So, that's… I think that's useful. It's almost meta in its way. Step back from the plan when it thinks the plan is the first step, but there’s, sort of, a recursive aspect here. Evaluate the plan for these various metrics and qualities, and that'll give you an early sign as to whether you're about to take, you know, a step down the right path.

29:25.97

Chuck Melley

Right, right. Yes. 

29:28.18

Eric Stano

Terrific. You know, a lot of the listeners of this podcast are in curriculum and product development, and we already referenced the fact that those are on much longer cycles traditionally than AI product cycles certainly are. And I wonder, having worked at an education company like mine, do you have any advice for education companies should decide what part of the AI learning experience needs to be frequently attended to by them in the products that they develop, in the content that they put forward, and, you know, or in acknowledging, you know, certain aspects that remain stable. I'm just wondering, you know, in this whole calculus, how should educational publishers and tech companies, curriculum providers, product development players, how should they approach this, and how can they be the most helpful in this context?

30:28.41

Chuck Melley

Yeah, what I'm thinking is there's like a consistent core that will stay the same, and then there's the modular, updatable layer right on top of it.

30:37.99

Eric Stano

Sure.

30:38.41

Chuck Melley

And what's so terrific about instruction materials and curriculum are delivered now; it's all electronic. So, we don't have to wait three years for another edition to come out, right?

30:50.72

Eric Stano

Right.

30:51.35

Chuck Melley

And so, it can be updated, you know, on a regular, you know, relatively regular basis, right? So, as far as the consistent core, you know, those kind of, touched on the things I touched on with all those letters in the framework, you know. And it's the concepts really; it's the data, the algorithms, the prediction, the, you know, the bias evaluation, you know, and keeping all of that… that is going to be the constant. But the modular, you know, datable layer is more about examples, you know. Like up-to-date examples with the current tools and, you know, activities that, you know, that reference current, you know, products and what's going on right now. But the thing is that has to be updated relatively regularly, you know, as far as that second part. And when that is done in such a way, I think it's focusing on what I maybe mentioned a little bit a while ago: we have to make sure that we don't focus on the tool, you know. That the modular, you know, updatable layer is about examples, you know. It's because we can't really teach the tools because the tool is going to change.

32:01.29

Eric Stano

Right, right. Makes a lot of sense. So Chuck, now we turn to a part of this conversation that my team has titled the lightning round. So, what I'm gonna do here is just fire a few questions your way and just ask for, again, a quick, off-the-cuff response. So, forgive the pacing here, but here it comes. Good. Name one AI skill schools are currently overemphasizing.

32:36.64

Chuck Melley

Prompt writing. I mean, it's useful, and it's great. And, you know, I try to improve my prompt writing every single day. But it's one of the, if not the most perishable skills, you know, on the list that we have.

32:51.54

Eric Stano

Very nice. One foundational concept every student should understand about AI.

32:57.94

Chuck Melley

Oh, yeah, this is the one I think I talked about with kids in elementary school.

33:02.78

Eric Stano

Yeah.

33:03.68

Chuck Melley

It's made by people.

33:03.56

Eric Stano

Yeah.

33:05.47

Chuck Melley

You know what I mean?

33:05.70

Eric Stano

It's not magic.

33:06.90

Chuck Melley

It inherits our strengths and our mistakes, and you gotta check it. You know?

33:12.65

Eric Stano

Right. Working at a company called Magic, it's always hard for me to say the word " not magic because everything surely should be magic. But AI is not. It is made by humans, as you say. One AI policy mistake states should avoid.

33:31.62

Chuck Melley

Yeah, and that's going back to the, well, you know, I guess I can say tool. I want to say the tool, right? You know, if they're writing code around a specific tool, that's, you know, where the snapshot is. Let's just not say tool; let's say the snapshot. If the policy that they're writing today is based on the snapshot of today, it's not going to be durable at all moving forward. And the second one is just because, you know, no funding. It has to fund, especially funding for teacher professional development.

34:01.35

Eric Stano

Right. And I am going to read this aloud because it is on the piece of paper in front of me, but I think you just answered this. One sign that a district's AI literacy strategy exists only on paper.

34:15.22

Chuck Melley

Oh, right. Right. Yeah. Yeah. And you know, it's an acceptable use policy, which says you can't use it in between the bells or something like that. But with no teacher professional development and no time on the schedule to teach, you know, it's really interesting. You know, kids, learners, students really, you know, should, you know, be brought up to speed on AI and how it's made, created, and best utilized. And if we don't have the time dedicated to it, or, you know, the attention dedicated to it through teachers and professional development, it's not going to happen.

34:56.00

Eric Stano

Well, you surprised me there. I thought you were going to say money. But I appreciate i appreciate that the more nuanced answer.

35:03.35

Chuck Melley

I said money five times. 

35:05.32

Eric Stano

Yeah, I know. Fair enough. Fair enough.

35:07.70

Chuck Melley

Yeah.

35:08.16

Eric Stano

Last question. One thing EdTech providers misunderstand about preparing students for AI.

35:14.96

Chuck Melley

Yeah. Yeah. I mean, comfort isn't competence, right?

35:19.15

Eric Stano

Mmm.

35:19.21

Chuck Melley

You know, getting good at using a tool, you know, is not the same as understanding it. Sure, they…

35:25.40

Eric Stano

Amen.

35:25.94

Chuck Melley

… they could be the best, you know, prompt writers in the world, but they might not know how it works. And if they don't know how it works, and what's behind it, you know, that's what, you know, EdTech providers really should be focusing on: that latter piece.

35:42.13

Eric Stano

Now I always appreciate both an alliterative and substantive answer, so thanks for that. It's the old English major in me. So we're about to wrap up, but this is the point at which, you know, I ask you to share, really, some advice or some tips, if you could, for the listeners. And there's just a few questions I'll ask along those lines, you know, for state and district leaders who might be listening. What's one decision they should make this year that, you know, to move from managing AI use to building genuine AI literacy?

36:25.04

Chuck Melley

Yeah, I mean, they should really implement policies. You know, I mentioned the 10 policies that I touched on before, and they don't have to implement all 10 of them, but even the, you know, one or two of them to get the ball rolling. They have to recognize the difference between, you know, AI use and AI literacy. It really… when they think about policy development, they focus on moving students from, you know, consumers to users to creators, and then, maybe, shapers, as I, kind of, touched on before. 

36:59.15

Eric Stano

Great. And for curriculum and EdTech product leaders and providers, as we were just talking about a few moments ago, what's one commitment they should make so that the learning experiences they build and sell to classrooms, you know. What's the one commitment they should make to make sure those learning experiences remain useful even as the underlying tools that they're applied to change?

37:25.24

Chuck Melley

Yeah, you know, I think the one thing that they should keep in mind is to make sure that, you know, what they provide students can, you know, explain how the AI works, right? You know, and focus on and, you know, touch on them before. That they focus their efforts on successful outcomes, right?

37:36.41

Eric Stano

Right.

37:44.04

Chuck Melley

You know, for the students. And measurable as well.

37:43.94

Eric Stano

Right.

37:47.35

Eric Stano

All right. Well, Chuck, you know, I want to thank you for joining me today. And, you know, it's always a... It's always a treat for me to meet a fellow alum of a company that I worked at for so many years. And somebody who's also tied to the political apparatus in this country. Before we part ways, though, where could our listeners follow your work and learn more about the policy and AI literacy issues that you're tracking?

38:18.48

Chuck Melley

Oh, definitely. If they go to Code, we still have it at code.org. That's our CodeAI website. That is where our 10 policies are listed, along with a great deal of other resources, as well as links to the AI Literacy Framework and a bunch of materials, and actually our classes as well, and our products. And then, also, they could reach out to me if they like,  you know, to write to CodeAI. But one thing, before we go, Eric. I wanted to touch on… I want to remind everybody that, in 2029, PISA is going to... PISA, you know, the International Assessment PISA. They're going to have an AI literacy assessment in 2029. So the clock is ticking, you know, for countries, for states, for, you know, municipalities. That this is going to be measured in 2029, and so they recognize that this is important. So I think, you know, it'll be really interesting to see where we are as a nation, and where countries are; the world is overall on this topic, you know. Once we get the results back in, I guess, probably four years- when we're going to get the results.

39:40.22

Eric Stano

Well, now that you've communicated that the clock is ticking and induced heightened anxiety in all of our listeners, that seems like the perfect place to conclude. I'm just teasing. But I want to again thank you, Chuck Melley, Senior VP of Policy, Advocacy, and Communities at CodeAI, for joining me today. Our listeners can follow Chuck as he's just described. And I want to thank everybody for listening today. I'm Eric Stano, and this is Tech in EdTech.