Welcome back to Teachers in America, where we connect with real educators and ed leaders to provide practical instructional tips and talk about the latest teaching trends to help you stay on the forefront of what’s new in education.
Today we are joined by Peninsula School District Chief Information Officer Kris Hagel and Chief Academic Officer John Yellowlees, who are using artificial intelligence and Universal Design for Learning (UDL) to foster an accessible education for all students. This past summer, Kris and John lead a presentation on their innovative approach to AI at the Model Schools Conference. In this episode, they sat down with Center of Model Schools managing partner Dr. Joshua Starr to discuss how to integrate AI into classroom lessons. Plus, they shared leading principles and strategies to guide students in using AI ethically.
A full transcript of the episode appears below; it has been edited for clarity.
You can follow Teachers in America wherever you listen to podcasts. Subscribe on Apple Podcasts, Spotify, Google Podcasts, Stitcher, or iHeartRadio.
Please consider rating, reviewing, and sharing Teachers in America with your network. We value our listeners' support and feedback. Email us at shaped@hmhco.com.
The views expressed in this podcast are those of the guest and do not necessarily represent those of HMH.
Jenn Corujo: From paper and pencil to WIFI and AI, education is ever evolving. On this new season of Teachers in America, we’ll keep you on the forefront of what’s new. We connect with teachers and ed leaders to talk trending topics and real issues, bringing you inspiring ideas that will influence the future of your teaching.
Today we’re excited to welcome Dr. Josh Starr to the Teachers in America team. Josh comes with a leadership background, having served as a superintendent of schools, and is currently a managing partner for the Center for Model Schools. This past summer at the Model Schools Conference, Josh had a chance to connect with Peninsula School District leaders, Chief Information Officer Kris Hagel and Chief Academic Officer John Yellowlees, who together are making education more accessible using artificial intelligence. In this episode, they’ll discuss how to integrate AI into the classroom and share strategies for using it ethically.
Now here’s Josh, Kris, and John!
Dr. Josh Starr: So, John and Kris, thank you for joining me. First, how are you enjoying the conference?
Kris Hagel: I would say the strong focus on pedagogy in all of the sessions and throughout has been really impressive to see so far.
Josh: Yeah, that’s great. And that’s what we try to do is focus on teaching and learning. So, this is where I think that you have something to teach us about how you can use AI to actually improve teaching and learning. Because what I’ve been seeing in the conversation oftentimes is people may be a little afraid of it, that it’s going to somehow hurt teaching and learning. But you all have taken a different approach. This could actually elevate and improve teaching and learning to take to the next level. So, talk about what you’ve been doing in the district.
Kris: Sure. We actually had teachers that were working through some learning around AI even before ChatGPT got out. In the spring of ‘22, we had a cohort of teachers that were taking some facilitated professional learning through ISTE on some of the AI tools that were available at that time. We look back at them now and they were pretty rudimentary at the time, but they really seemed like they were the forefront.
Josh: What kind of tools?
Kris: Just some very, very basic text generation or image creation tools, more on the machine learning side of artificial intelligence. Machines can learn from the data that’s out there. It was really just to dip our toe in to see what’s going on, to make sure that we’re staying in front of what might be coming. But at that time, I think we thought we were looking at something that was going to impact us five years down the road, not three months down the road.
Josh: You had teachers that sort of started doing it. So, has this been a teacher-led initiative or just in collaboration with teachers? Because it’s always powerful I think, when leaders at the central office level are saying, “Hey, let’s actually see what real-life teachers in real-life classrooms are doing.” So, talk a little bit about that process of how you’ve gotten teachers involved.
Kris: Sure. We’re kind of in an interesting place where, unfortunately for us in our district, we had to make some significant budget cuts last year. We lost a lot of our central office instructional facilitation. A lot of those people ended up put back in the classroom because we needed to cut about 8% of our budget. We’ve relied really on this being a teacher-led initiative. I mean, we steer it and guide it in a certain direction, but for the most part, it’s taken on by a group of 30 to 35 teachers in our district that really are passionate about it and see the future. And then we just kind of help make sure it goes in the right direction as they start talking and developing and thinking through what this looks like in the classroom.
Josh: Yeah. It’s interesting how the crisis of a budget cut can create an opportunity, right? So, John, as the CIO, if Kris is starting to come to you and saying, “Hey, teachers are starting this interesting stuff.” And you’ve got to report to the superintendent and the board and you’ve got to think about, okay, let’s bring some sense in order to this so that people aren’t just doing whatever they want with this new tool that we still don’t really know what it’s going to do. How did you then approach how AI should fit into the larger environment of the school system?
John Yellowlees: Well, I think it really fits nicely with our work around UDL. We’ve had an initiative for the past six years, I believe.
Josh: And for UDL? Can you talk a little bit about what UDL is?
John: UDL is Universal Design for Learning. It’s really providing students multiple means to access content, to engage in the content, represent their learning. As part of a broader MTSS initiative that we started about six years ago, we started to weave in UDL as really a means for all students to access content—grade level content—in a meaningful way. The more we were working on UDL and as AI emerged as a potential tool to help out and provide teachers with that, that’s kind of where the integration occurred and where the connection with teaching and learning really took place.
Josh: And what was the response? You’ve got your sort of leading-edge teachers who are saying, “Hey, we’ve got to do some stuff with this guys. Let’s make it happen.” Then you’ve got others who are probably sitting there like, "I can’t deal with one more thing." What was the response? You started integrating this with UDL and started rolling things out?
John: I think there’s definitely, as in all cases, some early adopters, and those that are a little more reluctant and kind of “this too shall pass” kind of mentality. And as we began that work this time last year really was in engaging in professional learning with our teaching staff. And as we began. That work this time last year really was en engaging in professional learning with our teaching staff. I think that prior connection, that prior work we'd done around UDL, um, and some of the resources that we've since provided to teachers that are pretty accessible for educational purposes around ai.
Josh: How do you deal with quality control?
Kris: Yeah, so that’s one of the things, and I don’t want to speak for John’s whole team, but I know I talked to them quite a bit and those conversations we’ve been having . . . is this just another problem that we’re generating for ourselves now with all of this content that’s not as quality?
It’s a little different though because teachers are actually targeting towards our curriculum, targeting towards the standards that we need to be meeting, not just downloading the cool worksheet out there. And so we have some more work to do and I know that his team is working on that, and how do we vet some of this content?
John: I think Kris and his team have done some great work in professional learning around prompt engineering. But I think regardless of what tool we’re using, we’ve stressed the importance in our teacher group, they’re doing a great job of whatever AI might produce, sitting down, spending time reviewing that. So say, for example, if we’re utilizing a different tool to generate a behavior plan for a student, for example, the prompt engineering piece and what we’re putting into AI is critical, but then sitting down and reviewing what that spits out in terms of, in this case, a behavior plan with the entire team and those who have expertise around the table is critical.
Josh: Right. As I’ve been diving into AI a little more and reading Ethan Mollick stuff and Ezra Klein did some great podcasts. That kind of introduced me to some of it. This notion of prompt engineering, because it seems like the intentionality and deliberateness of understanding the question you want to ask, that problem you’re trying to solve, is, to me, what seems to be the real power of AI.
Not just the ability to spit information out. I’m curious how you actually teach people, how you teach adults that. And I’m also curious about the application for kids.
Kris: There’s been some good work out on this. I took and transformed a course that a professor from Vanderbilt put together on prompt engineering. And I took that. His course was 10 hours long, and I took it and brought it down to four hours for our cabinet, and we did that for adults. And then one of our teachers that have been really passionate about AI, he took my work and then moved it down to a 45-minute lesson that you could do in a staff meeting when the principal gave him 45 minutes.
It’s a skill and we’re having a lot of conversations about, "is this something that’s a skill that we’re going to need for a long time or is this a skill that we’re going to need for a little time? As AI gets better, is this still going to be something we need?" For right now, it sure looks like it. So we’re trying to teach it to teachers who are trying to then teach it to staff. There’s some great resources out there. There’s organizations that are putting out some really nice frameworks around writing good prompts, easy for teachers to understand, to follow, and figure out how to write a good prompt to get a good response back.
Josh: One of the things I’ve been thinking about when the coding craze started, 15 years ago and coding could be the next coal mining. I think about that with prompt engineer. How do you keep people understanding that this is going to be generative, right? It’s going to constantly be new. We’ve got to constantly be on top of it.
Kris: There’s a conversation going on right now that I think is really going to be fascinating to watch over the next year. And that’s whether, I guess to say it the easiest way, is the next coding language going to be English? Because that's what it seems that it's going towards. Maybe it’s not the nuts and bolts of where to put colons and commas and everything in a programming language, but how to really say something in very clear and detailed language to get AI to do what you want it to do for you.
Josh: So totally different kind of literacy.
Kris: Yeah.
Josh: If that’s the case, this is where I’m curious from the CAO perspective, when you think about more broadly, how do you start integrating that kind of thinking into teaching and learning and curriculum for kids, because they’re going to be the ones that are doing it in 2035 and 2040 and 2050. And they’re going to be doing it in your region in the Seattle area where you’ve got a lot of tech. So how do you teach kids to start thinking that way?
John: Well, and you can chime in on this too, Kris, but I think on a smaller scale, that’s what some of our teachers have been doing. We’ve involved some teachers in some professional learning over the past year, year and a half. And some of the same modules that Kris was talking about in terms of those that we administered to our leadership team, principals and others, teachers are then taking that prompt engineering piece and working with their kids.
We have a teacher at one of our middle schools, for example, who is working with kids and students on that prompt engineering piece. What do you put into ChatGPT to get the response that you’re looking for and really kind of analyzing that and working through that with students. Not just the final product, but the work that leads up to that final product, say for a social studies assignment that might be administered.
On a small scale, our teachers are working on that. Those that we’ve been meeting with are really pioneers in this field, but have talked with us quite a bit about how do we expand this more broadly so that all of our staff are getting the same professional learning. That time for professional learning and expanding that training is really kind of the challenge that we face now.
Josh: What’s been the response from kids? And I’m also curious about the response from parents too, if you’ve seen anything.
John: Yeah, Kris can speak to that. He gathered some great survey data last year on that very topic.
Kris: When we started bringing teachers together last year, we really wanted to collect data and get feedback from the teachers on how this is going and with the kids on how it’s going.
Kids have, of course, they have kind of an excitement, but also a hesitancy. We’re also starting to see more this year some of the kids really being thoughtful about, "I really want the things that I write to make sure that they sound like me, and I don’t want to be the one that’s just doing this because it’s the easy way out." So you kind of get both sides of that coin. Some kids really just want to get their project done. There’s others that are really becoming a little bit more resistant to using AI in some cases because they want to make sure that their voice is heard and that what they say and what they turn in . . . a lot of kids last year were thinking, "I’m just going to use this on my college applications. I’m going to use this for everything that I just need to turn in and get it over with." And now kids are kind of rolling that back and saying, "No, I want to do a better job than what I’m going to get."
Josh: That’s fascinating, right? I want to make sure I get this right. So, the assumptions that we as adults have about how kids might use it, [like]cheating. You’re actually seeing something different? I think it’s really fascinating. Do you have student voice in how you think about all this? Do you engage kids in it? I think it’s oftentimes as adults our assumptions about how kids are actually going to respond and what they’re thinking may be totally different. How do you get them activated in helping you think this through?
Kris: We did a lot of survey data last year, and then this year we’ve kind of done more on the focus group side. We’ve talked about AI, our superintendent has a student advisory council. We’ve done it there. I’ve been in meeting with leadership team, our teachers, and their kids.
We also have another group of kids that are going through a teacher apprenticeship program at the high school. And I sit down with them regularly because their teacher’s big into just making sure they’re prepared with the AI tools that they’re going to need moving forward. And so I have a lot of conversations with the kids. That’s how we get some of this feedback around how are they feeling and what’s going on.
John: Yeah. Done some real intentional work at one of our high schools, our alternative high school, this year with AI and UDL. And I think like Kris was speaking to, that rather than surveying that, sitting down with students, that focus group type environment, you get a lot of really powerful feedback from that rather than sending out a survey. We utilize that as we move forward.
Josh: What was most surprising that the kids said to you in those?
John: I think like Kris was alluding to, just that desire to have their own voice come through, rather than it be something generated for them on behalf of them. We’ve done some other innovative work in the district around cell phone usage and or lack thereof, and we make assumptions as adults that this is not what students want. But when you actually sit down with kids, whether it’s the AI or the cell phone piece, the reports we get back are, "This is great. Not having cell phones in the classroom has allowed us to engage more with our peers and attend to instruction." And the same is true with the conversations we’ve had around AI.
Josh: Yeah. It’s amazing how kids end up rising to a level when they’re engaged and interested and all that. So let’s talk about scaling a little bit. Let’s talk about different applications. Particularly getting to the idea that you’ve got a lot of different kinds of kids. They come at, and this is the UDL piece, but talk more about how you’re thinking about scaling and particularly meeting the needs of diverse learners, whether they’re neurodiverse, whether they’re speaking another language, whatever the case may be.
That’s one of the challenges. We want to make sure these kinds of tools are equally accessible to everybody. How are you approaching that with integrating AI?
Kris: I think the easiest way to start with that is probably from the lesson planning standpoint.
One of the things with UDL that we heard—as you know, we’re getting teachers up to speed and getting them implementing UDL practices in the classroom—is that it takes a lot of work. It takes a lot of time to build these lessons that work for everybody. Utilizing AI you can do that much faster, to do more personalization. We started working on some training with some of our teachers this year on building their own AI tools to help them personalize really down to the student level. Taking preferences from the kids on how they like to learn, what are the mediums that they like to use to learn with, and then having teachers be able to design lessons specifically for those students. There’s some really positive work going there.
Josh: When I think about differentiation, some of the old model. . . Or, you know, when I say old, the current model is . . . well, take the unit assessment test or the growth measure, do the item analysis, look at it. There’s this question, what kind of data you would use to differentiate lessons, other than student preference. What kinds of data gets put in to that conversation so that the teacher is then using the tool to differentiate appropriately?
Kris: One of the things that we’re still unsure on is some of the data privacy pieces and how you do this safely. We’re trying to do everything we can still do it in a manner. We’re working on some tools that we think we’ve figured out most of the data privacy pieces and those will roll out next year. For this past year, what we’ve been doing is things like using nicknames and tying it to data that kids have or not putting in student numbers and their screening scores or any of those types of things, because we’re still not comfortable doing that in any safe environment. I think we will be next year, and I think it’s the future of what we’re going to see moving forward. We’re really working on what are the kinds of things. . . Right now it’s more of do you prefer teacher lecture? Do you prefer reading this? Do you prefer watching a video? You know, those kinds of less-sensitive pieces of information as you’re building out lessons.
John: I think in general, we know that every classroom there’s a wide range of variability, so it’s more of that proactive. So differentiating in the moment based on the students I have in front of me. UDL is really that more proactive approach to say that here’s the standard we’re working on today, and before I start the lesson, here’s a number of ways that kids can engage in the learning.
There’s a wide range of options for students to engage in the content that’s related to the standard. Then a range of options for them to demonstrate their learning. So it’s not the traditional sit and get, I’ll do some differentiation, then everyone will produce an essay at the end.
Josh: That’s interesting. And I do think that the next level is going to be the data privacy issues. To then connect it to some of the quantitative data we have. But you’ll be well prepared for that, it sounds like. So part of what I’m also really curious about is the . . . there’s a change management process going on here.
I’m really curious about that. And I also want to talk about the policy environment. So first talk about how you’ve approached this from a change management perspective to be seen as something that you’re leaning into. Not resisting and is not just like one more thing, but it’s like, "Hey, this actually going to help us solve it." How have you led the change management aspect? Then I want to talk a little bit about policies.
John: I think that integration is key. I think to your point of seeing it as one more thing as we started our MTSS work, we talked about this a little bit in our presentation today.
Six years ago, UDL wasn’t really something we were focused in on as part of our MTSS framework. As we made that, and as we learned more about that and integrated it, and it’s kind of now a foundation of our MTSS work and teaching and learning. As AI became more prevalent, we decided this cannot be another thing in addition to what we’re doing.
It really has to be integrated. So we’re integrating it into the work that we do. We’re just planning our administrative retreat for August and rather than have standalone AI learning, we will integrate that into everything that we do with our principals and expect that our principals integrate that into everything that they do.
Josh: So it becomes an accelerant almost. I’m really curious about that with leaders. Because I’ve been wondering about that at the leadership level, what kind of AI tools we should be looking at to help adults do their work better? Have you all been thinking about that?
Kris: Yeah, a lot. We could do a whole other podcast on that. We’re doing a lot of operational work around that right now too. One of the big things, when you start really digging into how do you going to make change across your entire organization and utilize AI is how do you prepare all of your data across the entire organization to make it accessible? There’s a lot of work that we’re doing now, specifically operational departments moving things in the right places so that we can build upon that.
It’s a big change and one you have to lead everywhere throughout the entire organization. One of the things I wanted to talk about as far as change is concerned, we have taken this AI and we’ve looked at it a little differently than we historically have. Some of that we alluded to earlier with the financial constraints we’re in right now. Normally we would say, "All right, this is what we’re doing. We’re going to bring in speakers. We’re going to get people trained. We’re going to do that." We don’t have that ability right now. And so we’re trying to figure out how to lead from, essentially lead with teachers being passionate about it. And, you know, we’re seeing huge uptake.
Josh: Wow.
Kris: Because we also have this focus with our teachers’ association around teacher wellness this year. What can we do to reduce workload, reduce burnout, reduce stress? Some of these AI tools can really help teachers save a lot of the time of some of the minutia that they have to do for their jobs. I think that’s one of the huge reasons why teachers are embracing it so much is because they can see the actual reduction in some of the things they have to do.
Josh: And the fact that they’re leading it too is so powerful. Because what we’ve seen. And I won’t go deep into, into James C. Scott’s book Seeing Like a State, but there are all of these examples of how people at the top think that something’s going to work and the people at the ground are like, "This isn’t going to work."And when you engage them, you actually design a much better solution.
So let’s talk about policy for a minute, because that is something that when I talk to superintendents and CAOs around the country and I say, "What are you doing with AI?" They’re like, "Well, we’re not letting kids use ChatGPT. No one wants us to cheat." You’ve taken a totally different approach. Talk a little bit about how you constructed kind of a policy environment and guidelines around this.
John: I’ll let Kris speak to that. I mean, I think a big focus for us from the teaching and learning standpoint has been the ethical use of AI. Kris was kind of a leader in that respect, in terms of generating some philosophy and belief statements around AI last year.
Josh: So, philosophy and belief statements. Talk about that.
Kris: You know, we had pulled teachers out of the classroom a couple times last spring, and they said, "You know what? We really need the district to come together and say something going in next year about what do you believe AI should be in the classroom."
Josh: So, like your values.
Kris: Yeah. Right. And so, like any good technologist I took. . . there wasn’t a lot written at the time.
Josh: So you asked ChatGPT.
Kris: I absolutely did, and I’ll tell you how. Right then it was about the time when the U.S. Department of Education came out with their Teaching and Learning with AI document. COSIN and the Council of Great City Schools had put out another document on AI and education. I read through those, picked out the highlights that were important, and then I took all of the highlights and put it in ChatGPT and said, "Come up with a rough draft of this. As far as the school district put a belief statement around AI in the classroom."
We then took that rough draft and ran it through teaching and learning staff and then back to the teachers and said, "What do you think?" Everybody made some edits and that’s what we published this year. And then on our website you can actually see the citations of the conversations from ChatGPT to how we got to where we are. So, you can follow along the whole thing. And that has led to a lot of other assistance we’ve been providing both in our state and nationally. That principles and belief statement is used in, I think, at least 10 different states across the country.
Josh: It’s powerful. I’ve seen it. It’s awesome.
Kris: Yeah. I keep getting Google alerts from other districts that are just crediting us and adopting it.
Josh: What I love about your story though is that you started with your own shared values. And to say, "Okay, here’s the shared values of our district and here’s our beliefs about what it should be," and then you build the policy on top of it. That’s really powerful. So, if people are just taking it from you rather than doing it on their own, it may actually miss the mark, frankly.
Kris: Yeah, it might.
Josh: But it’s still something.
Kris: There are a lot of superintendents that will call us and say, "Hey, where do we start?" And I’m like, "If nothing else, just take the one off our website and take that one."
Josh: You should say, "Well come see us speak at Model Schools."
Kris: There you go. That’s it.
Josh: We’re going to wrap up. I just have one question and then one bonus question. Well, actually two quick questions. What’s been the biggest surprise or learning that you’ve had, like a realization that you’ve had in doing this. And what is a quick advice you have for others, other than just download your principles. What’s been the biggest surprise and what advice do you have for others? And then I have a quick bonus question for you.
John: I think the biggest surprise. . . I did not realize how effective this would be in terms of, as we started our UDL work, really integrating with UDL. And we’ve just seen teachers and, I think primarily of some of our special education teachers, really utilize this as a way to be inclusive and create more inclusive learning environments working alongside general education teachers for our students with disabilities. And that’s happened within a 12-month period. That has surprised me, a pleasant surprise.
Josh: Alright, Kris, what’s been the biggest surprise?
Kris: I think teacher adoption really. Once they see just a little bit of how it can help, how deeply they’ll jump in and dig into it. Once you open the door and say, "Yeah we want you to do this, feel free to explore and figure this out." And how quickly and deeply our teachers will embrace it and figure out what’s going to work in their classroom. They really appreciate being a part of the conversation as we’re building it, because nobody really knows what two weeks down the road or two months down the road or two years down the road is going to be. But if you’re digging in now and taking part now, you can help build what that’s going to look like two years from now. So you can be a part of those conversations. And so I think our teachers really have embraced that moving forward.
Josh: What’s your one word of advice for folks who are dipping their toe in the AI water?
Kris: Get started. You got to get started. There’s still so many people that call and want to know how to get started, and there’s no roadmap. Just try something. You’re not going to hurt anybody if you don’t get it right the first time.
Josh: You learn the work by doing the work.
John: Yeah. I’d say the same thing. Don’t be afraid to get started. You’ll make mistakes and learn from those and continue to build capacity, but you have to start somewhere.
Josh: Alright, quick bonus question. What is the most joyous thing you saw in schools this year?
Kris: I would say we just did a tour with our Washington Association of School Administrators, one of their leaders who’s going to be leading AI work. And we toured a bunch of classrooms and just the joy on kids’ faces when something clicks for them that has never clicked before, be that in a high school resource classroom or even in an elementary classroom. Just things that have never really made sense. The use of AI has helped them to just finally have something, just finally click and go, "Oh, yeah, I get it now."
Josh: It’s just that light bulb. What about you?
John: It’s hard to choose. With the end of the school year, it’s been a couple weeks of really being not in buildings and then the year graduations and celebrations, and I’m thinking to just seeing some of the highlights from some of our schools. At one of our elementary schools, a group of students with their teacher released a balloon up in the atmosphere. And just seeing the kids out on the field and the celebration around that and the curiosity. There was that paired with it’s the last week of school and field days. So just a great way to wrap up the school year.
Josh: Yeah, those are fun. Well, look, Kris, John, thank you for joining me today in this episode of Teachers in America. Great conversation. A lot for people to learn. I love it. Just get started, right? Just start doing it. Sounds like you all, I know you’re all doing really cutting-edge work, so I appreciate you sharing it with us, and glad to see you at the conference and we’ll keep talking. So thanks for joining me today.
Kris: Thank you.
John: Thank you. Appreciate it.
Jenn: If you or some you know, would like to be a guest on the Teachers in America podcast, please email us at shaped@hmhco.com. Be the first to hear new episodes of Teachers in America by subscribing on Apple Podcast, Spotify, or wherever you get your podcasts. If you enjoyed today’s show, please rate, review, and share it with your network. You can find the transcript of this episode on our Shaped blog by visiting hmhco.com/shaped. The link is in the show notes.
The Teachers in America podcast is a production of HMH. Thank you to the production team of Christine Condon, Tim Lee, Jennifer Corujo, Mio Frye, Thomas Velazquez. and Matt Howell. Thanks again for listening.
***
Hear new episodes of Teachers in America on Shaped, Apple Podcasts, Spotify, or wherever you listen to podcasts.
Hear more of Teachers in America on Shaped
Related Reading
Photo courtesy of Leeping Lenses Photography.
