Show Notes
AI can make a task easier without necessarily making learning stronger. The distinction may come down to who is doing the cognitive work.
When AI provides an answer or produces the work for a student, it can reduce the attention, engagement, reasoning, and productive struggle involved in learning. When thoughtfully designed and intentionally used to provide hints, scaffolding, practice, or feedback, however, AI may help students remain actively involved in that process.
For educators, the goal isn't simply deciding whether to use AI. It's considering how the tool is being used and whether students are still doing the thinking that builds knowledge.
Books & Resources Mentioned in this Episode
Stanford SCALE — AI Hub in Education
The hosts reference research and resources from Stanford University examining the emerging evidence around artificial intelligence in K–12 education. The review discussed in the episode highlights how limited the current evidence base remains and why educators should be cautious about making broad conclusions from early research.
Tim Odegard and Megan Gierka — Using AI to Support Structured Literacy –Aligning Tools with How Students Learn
Stacy references an article by Tim Odegard and Megan Gierka that considers AI through the lens of the instructional hierarchy and explores ways technology might provide students with additional opportunities to respond, practice, and receive feedback.
Big Sky Literacy Summit
Several ideas discussed in the episode connect back to presentations and resources shared at the Big Sky Literacy Summit, including conversations about AI, learning, and the distinction between having access to information and developing knowledge.
💬 Want more insights like this?
Subscribe to the Literacy Talks Podcast Digest for episode recaps, resources, and teaching takeaways delivered straight to your inbox!
Core Instruction and Reading Intervention. One System. One Path Forward.
From Pre-K readiness to advanced fluency, Ascend®️ gives educators a consistent, needs-based approach to literacy—helping every teacher deliver the right instruction and every student build lasting mastery, one skill at a time. Coherent literacy for lasting impact.
Do you teach Structured Literacy in a K–3 setting?
Sign up for a free license of Reading Horizons Discovery® LIVE and start teaching right away—no setup, no hassle. Sign-up Now.
View Transcript
Narrator:
Welcome to Literacy Talks, the podcast for literacy leaders and champions everywhere. Brought to you by Reading Horizons. Literacy Talks is the place to discover new ideas, trends, insights, and practical strategies for helping all learners reach reading proficiency. Our hosts are Stacy Hurst, a professor at Southern Utah University and chief academic advisor for Reading Horizons; Donell Pons, a recognized expert and advocate in literacy, dyslexia, and special education; and Lindsay Kemeny, an elementary classroom teacher, author, and speaker. Now let's talk literacy.
Stacy Hurst:
Welcome to this episode of Literacy Talks. I'm Stacy Hurst, and I'm joined by Donell Pons and Lindsay Kemeny, and we are talking all things literacy. And today we have a big topic, and it's kind of a hot topic, and I am so glad that Donell is going to be leading our conversation because there's so much to this topic. So I'll turn the time over to you. Introduce our topic and and let's have a fun conversation.
Donell Pons:
On that note, let's see if I can make this happen. But yeah, Stacy has told us this is going to be about AI, and it's hot, hot, hot. I'm sure we'll have many episodes about it, but here's where we are so far. And part of the reason to do the episode was to share some great resource. We were at the Big Sky Literacy Conference, and there were some resources shared from Stanford University. So that's what we're going to be referencing here. Is a lot of information from Stanford, who's trying to to create a repository so that you can get information on what research we have and what are some of the culminating thoughts about AI in education. So I just thought I'd throw that out there for everybody to know because I was unaware about how much they had and appreciated knowing that. The piece we're going to talk about is a sort of report and research piece, and its key highlights from the evidence base on AI in K through 12, a 2026 review, and just to set this up before we start discussing pieces of it, what they did is they looked at about 818 papers that were related to artificial intelligence in K through 12 education. Glad they did it, and we didn't have to. But what happened was, is looking at the 818, they decided to come up with okay, but we have to have some parameters, and they have to meet certain guidelines. And when they got done with the review, at the end they only had 20 papers that met the criteria for strong causal evidence, and that's important. And the researchers also found that much of the existing research was conducted outside the U.S. in the K through 12 classrooms and involved short-term outcomes, and many had not yet undergone peer review. So, just to make sure you're aware of those things. So, again, very little. I mean, we're in these early stages of this, but it's it's moving so rapidly that obviously we're all trying to jump on board and and figure out what we can, and for educators this matters because conversations about AI, like we're saying, are moving so much faster than the evidence that we have, and so you have to ask yourself, and that's what we're kind of going to do today: is what do we actually know, where is the evidence, and how is it limited, and how can we use AI without losing sight of what students need to learn independently? I think these are some of those big overarching questions that we have, and then again, fundamentally, does AI help students learn, or does it simply help them complete tasks? That's a really good question, and so let's just dive right in. Number one, I just told you is the evidence base is much thinner than people assume, and I just gave you that piece there. And then number two is AI can improve performance without producing learning. That's one of the findings. And so one of the things I was going to ask you guys to kind of chime in on here is if AI helps a student produce a polished response. How do we know whether the student has developed the reading comprehension, vocabulary, reasoning, and writing skills needed to produce that response independently? And does it matter? What do you think, Stacy? Do you want to go first? We have Lindsay a minute.
Stacy Hurst:
Yeah, I'm going to guess that Lindsay's students. Not too many first graders are using AI. I would hope. That's my natural lindsay. I know. I do encounter this on a daily basis in my position currently, and I have put into practice a couple of things. I have assignments that I say that students shouldn't use AI on, or I will be very specific about how to use AI. That being said, I know there are students who disregard that and just do it anyway, and they do outsource their thinking. So I've done a couple of things. I actually at the end of last semester, I gave a paper. Pencil test of the same questions that they've been asked in quizzes and in other assignments and places throughout the semester, and that was pretty telling. And I think for the most part, though, students were actually comprehending and learning the concepts. The other thing I've come to do, and I've learned that this is very beneficial, and I tell students at the beginning of the semester if I suspect AI because here's the other thing we don't have really good checkers, and when I say when I say use of AI, I'm not opposed to that, but I am opposed to it if they're outsourcing their thinking and like you're saying, Donell, like the question is asking if it's interfering with their learning, and then I will have a conversation with them, and they will not know what it's about. I'll just call them and say,"Hey, let can you talk for a minute? Let's you know da da. And a lot of times, if they've used AI, they have no recollection. I've got their paper or their assignment right in front of me, and I'm asking questions about it, and it becomes very apparent very quickly that they do not have any recollection of the content, and that has happened three times already this semester. So
Donell Pons:
interesting. So I'm hearing you say the process does matter.
Stacy Hurst:
Yeah, absolutely,
Donell Pons:
most definitely. And Lindsay, do tell us because we're dying to know. Are any of your little ones using
Lindsay Kemeny:
it? No, my first graders are definitely not using AI. So this is so funny though because this finding to me is like comical because I'm like, hey, I don't need a research study to tell me that, like, that's
Donell Pons:
not durable.
Lindsay Kemeny:
Yeah, and and when they're saying, you know, oh, the AI tools significantly improve student performance on math practice and writing tasks. Yeah, because AI is doing it for them, of course, of course. But yeah, and then like Stacy said, of course, if you are outsourcing the thinking, you're not doing the learning. So we have to be very careful. And really, I think if upper grade, you know, secondary and higher ed is going to use intentionally use AI somehow, then you have to be very intentional about the tasks that you're giving and having them do to make sure that they're still thinking and learning,
Donell Pons:
yeah, and I think it's interesting because Hugh Katz at the Big Sky, he did a presentation at Big Sky Reading Conference, and I think I've mentioned it before, but it has stuck with me when he said information is not the same as knowledge. I think we're really we're really letting that sink in. We have a great deal of information available to us. Perhaps some would argue too much, but what's our knowledge base like? That's different, huh?
Stacy Hurst:
Yeah, I think it does come down to the way we engage with it too. Because he also said, and well, he's not the only one to say this, but essential for learning is attention and engagement and interaction, you can, and I think AI just short circuits can can short circuit both of those things, especially yeah if you're not paying attention to what even AI puts out, right? You just submit it or use it as is, yeah.
Donell Pons:
And Lindsay, I'm glad to hear that at least in first grade, I should hope that space is being preserved. Yeah. Do you worry at all that you might start getting pushback about the skills you're teaching and people saying, "Yeah, do they really need to know some of this stuff? Because we'll just use AI. Have you heard anything like this?
Lindsay Kemeny:
I haven't heard anything like that, and I, I really doubt I will. I mean, we. I think everyone understands the importance of that. Those, you know, reading and writing basics. They need to be able to read and write. So, yeah,
Donell Pons:
yeah. Okay, interesting. It is really interesting to think about. Okay, the next big kind of thought that came out of this was the writing and reading finds findings deserve special attention, and so we already hit on this a little bit in what we were talking about. So what they found is that, and Stacy, this goes directly to what you said, and so might dwell here for a minute because it gives you a little stat to go with what you were saying. The report describes research in what students writing essays with an AI chatbot showed weaker recall, and that's part of what you were discussing. But here's the percentages. If folks were like,"Oh, geez, that's interesting, Stacy has that experience. It's not just Stacy. 83% of AI-assisted participants could not provide a quote from their essay. 85% who were assisted with their essay couldn't provide a quote from their essay. Now this goes back. How important is it for the recall? What do you think? Any thoughts, Lindsay? You,
Lindsay Kemeny:
I mean, when you're writing, you learn more. You have to grapple with the information and figure out how to like interpret that. You have to make sense. You have to make. It makes sense in your head, and then be able to convey that to someone else. So I think that a lot because as an author, I have
to think about:
I need to take this research, I need to make sense of it, I need to share how I apply that in my classroom. And AI doesn't know all that. And if I was using AI, like I like I just think the process of writing a book, I learn so much because you do. You have to grapple with that information and make sense of it. If I were just to ask AI and have it write a book for me, I wouldn't be doing any learning, and then like Stacy said, I wouldn't be able to recall like what I even wrote. And they're they're just there there there is a lot of people who are using AI to like produce things, and oh, it just sickens me. Really, I know. Do
Stacy Hurst:
you? I'm laughing because can you imagine what a book signing would be like, or a like a training if you had just used AI to write a book and then people ask you questions? Yeah,
Lindsay Kemeny:
yeah. And you get this. Like, I'm in a lot of writer groups, so you get this a lot in like these fiction writing spaces where, and you can see them on Amazon. Like that, people are spitting out books every month because they're just having they're giving prompts to AI, and it's just you know everyone it's called AI slop. Everyone they have a name for it, so we want to be really careful. But just thinking about our students, we want them to go through that process of grappling with that information, making sense of it, and doing the output-not just the input, but the output-because that is so that is so important for the knowledge, like you said.
Donell Pons:
So let's hang out here for a second. What do you say to the parent who says, "I have a child with a reading and writing difficulty, and AI is making this actually doable for my student. What do we think about that? What would your response be?
Lindsay Kemeny:
Well, it's kind of like accommodations and interventions. Like we need them to have accommodations, but they still need their intervention. So to me, it's kind of like that. Donell, you have the great example at the beginning of the year having your students with dyslexia use AI to write an email to their teachers explaining the importance of their accommodations and maybe their IEP and doing a little introduction. That is where that is so wonderful to use AI, I think. But then in the assignment, you also want them to teach. So learn. You want them to learn. You need to teach them. So there needs to be a. But perhaps you know. It's kind of like you know when you have shortened assignments, or maybe in upper you're like everyone has to do a five paragraph essay, but this student does a two-paragraph essay or something. I wonder if that is where you know they do the two-paragraph essay themselves, and then for those other three paragraphs, they could use AI. I wonder if there's a balance there to meet you know the needs, the accommodation. Does that make sense? I
Donell Pons:
know, because see, that's a modification, and we've often made this conversation, right?
Lindsay Kemeny:
Yes,
Donell Pons:
right. And so, would that be acceptable to think of AI as maybe being a way to actually have a modification that might make sense? I think
Lindsay Kemeny:
as long as you still have some of it, I think where they're actually doing the work.
Donell Pons:
Yeah.
Stacy Hurst:
Well, and I think starting with them doing the work, it has to start with them, and then have AI engage. And I think that's one of the things that is detrimental to our profession because AI is progressing so quickly that we don't have time to really research, let alone look at the research and see what the best ways are to explicitly teach our students how to engage with AI, because they do need to be taught that if you all you can do is speak into a microphone, and and I'm not diminishing this because again it's capturing thinking. If you're speaking into a microphone and saying this is what I want to say in my essay about frogs, and you you can put it together in a verbal form, and then have AI you know organize it. You still did the work up front, right? And then you still have work to do because you're not just going to accept what AI does because I am here to tell you it is frequently wrong, frequently, and so you want to be able to read it and revise and continue to engage with it,
Lindsay Kemeny:
yeah, I think that. But I think it's important because, like, especially for our students with learning disabilities and learning challenges, that we explicitly teach them. Because I think you know, after school, when school is all done and they need to email an employer or something, they're going to be relying on those. You know, they'll probably rely on AI, and so if they have been taught how to use it and to check it, and the you know, I think that'd be helpful. But then I think we need to be careful with. Students who do not have any kind of learning challenge or been diagnosed with learning disability that they don't start over relying on AI because then that's that is going to cripple them because they won't have skills that they are capable of and they need.
Donell Pons:
Yeah. So assistance not a substitute.
Lindsay Kemeny:
Yeah,
Donell Pons:
yeah, right. Okay, I think there will be more to come, particularly on this one because it does impact a lot of education. Okay, hand in hand with this was on note taking. Note taking alone or combined with AI significantly outperformed AI only use for comprehension and retention when it came to note taking, does that surprise you at all?
Stacy Hurst:
No, that tracks. That tracks, right? And I have an experience that just happened two days ago in my class. We have been talking about literacy. Of course, that's what I teach, and we've been talking about things like Scarborough's reading rope, and this is the beginning literacy course, and one of my students came up to me afterwards and said, "Do you have images to go with some of these concepts? Now we do obviously for the reading rope and for other things, but there was some content on my slide like vocabulary acquisition that didn't really have a picture to go with it, right? And I said to the student that would be a great use of AI. You have your notes, put them into AI, and say, come up with pictures that will help me remember these concepts, and then you negotiate with AI to think to say if that is going to help you or not.
Donell Pons:
Yeah, if that's appropriate.
Narrator:
Teaching literacy shouldn't mean juggling multiple programs. Ascend Mastery by Reading Horizons brings it all together-a core, comprehensive pre-K through five literacy curriculum that connects word recognition, language comprehension, oral language, and writing, built on decades of proven foundational expertise. Ascend Mastery simplifies instruction and helps every student build lasting literacy gains. Ascend Mastery will be available for district and school implementation beginning with the 2026 school year. Visit readinghorizons.com/ascend to learn more and sign up for updates.
Donell Pons:
This is really interesting because I have a lot of students who get a note-taking accommodation because the spelling is just still still so challenging. They're spending all of their time trying to spell rather than listening to the lecture and really being able to take notes. But what's interesting is for all of those students that I've worked with who have that as accommodation, they know the value of engaging with those notes, and so one of the strategies, if I don't already offer it, a student will say to me, "I have note taking given to me as a an accommodation, but I don't get a great deal out of them. I know I'm not utilizing them like I should. Do you have any suggestions? So even the students themselves know that if it's just handed to me, that isn't terribly helpful. I'm not really engaging with it. So we come up with strategies. A lot of them are drawing pictures off to the side. You get what you can from this material. What is it that you synthesize? What is it that you'll remember? And they'll draw pictures or make a little note for themselves off to the side. But they realized they have to be engaged with that in some way, right?
Stacy Hurst:
Donelle, I am so glad you shared that, and I'm sure many other professors will be too, because in what that looks like in college is you have designated note takers, and I have one in that very same class we're talking about, and I think I probably have three students that that particular note taker is taking notes for, but beyond that, I honestly haven't thought about talking to those students about how to engage with those notes, and I'm hoping somebody is. But what a great idea! Because I know the note taker is getting a lot out of that. She is always the one with the ready, correct answer, right? And that makes sense because she's very good at at processing the content into notes, but yeah. Then what? Oh, I'm gonna. Yep, I'm gonna follow up on that, and I'm gonna see what's going on. Yeah, yeah.
Donell Pons:
So each of them adopted whatever process. Where some were like to think of it as a picture. Some of them like to have little bullet points. So I just need this information to be in a bullet point format. But they made a goal to make sure they interacted with those notes because they said it's so easy to just get the notes and not interact with them. Like the whole point was interacting with the notes. Yeah. Okay. And there were different
Stacy Hurst:
ways that they interacted. And and I and what I hear you saying, just clarifying, is that it might be different for each person, but it will be the ultimate
question:
Is how do I get the most out of these? Okay.
Donell Pons:
Yeah. How are you going to do it? Yeah.
Stacy Hurst:
That's awesome. Thanks.
Donell Pons:
Okay. Now we're coming to a really interesting point. I think perhaps one we'll want to spend most of our time on here. AI design matters more than simply giving students that. So here it gets into the nitty gritty of the report does not treat all AI tools as equivalent. It identifies a distinction between general purpose AI and tutoring oriented AI, and there's a difference. And that if you're not paying attention, you might not notice it, but it's a huge difference. In one study, students using a tutoring specific chatbot performed similarly to textbook users, while students using a general-purpose chatbot performed worse on a subsequent exam. So the type of AI mattered, how it was set up, and the interaction the students were having. So the tutoring-oriented AI, just to give folks a little a taste of what do you mean by that? Can provide hints and graduated support. Can encourage students to reason through a problem. Shows greater promise when designed around learning principles, and they're actually built in. Learning principles are built into design. Whereas general purpose AI is just it's going to give complete answers. It's going to reduce the cognitive workload. So let's look at that again. That's that little phrase that I gave you, just in case we were kind of lost in it. In one study, students using a tutoring-specific chatbot performed similarly to textbook users, while students using a general-purpose chat bot performed worse on a subsequent exam. So, what do you guys think about that? When we are intentional about the AI, again, is this an aha moment or not? What do you think?
Lindsay Kemeny:
I think the difference is, you know, the chat bot, the more targeted one, is walking them through the process, so it's teaching them versus just giving them the answer, right? So it's a little bit more mimicking a teacher helping the student work through a problem or something, so and it doesn't
Stacy Hurst:
take away that what we call desirable difficulties, right? Like you're not taking that off the the table. Yeah,
Lindsay Kemeny:
they still have to work. The student still has to do some work. They're not just going to say, "Here you go. Here's the answer.
Donell Pons:
Yeah, and that all important word engagement, right? Cognitive is providing more of them, yeah. And so I think there's promise in that right there that caught my eye because I thought, oh, there's promise there that if you intentionally build the tool to have some of those strategies and concepts that you use as a good educator, a good teacher, that that could provide a lift for, say, just having one individual in a classroom where you might have a great deal of need, right? And provide opportunities. So within that, they did pull out some really interesting things that they focused in on that were different with the tutoring-oriented AI, and it said there was explicit instruction. I think you might have mentioned that, Stacy. Scaffolding, gradual release of responsibility, productive struggle, which was mentioned, and the need for students to engage in the cognitive work of learning those things. So, what kind of applications do you think we might have for something like that?
Stacy Hurst:
You know, I'm gonna I'm going to jump in here and say there is an article that I think we can link, written by Tim Odegard and Megan Gierke, and it actually takes the instructional hierarchy into consideration with tools like AI and how they could benefit. One really important thing for that transfer of learning is that students get enough response opportunities and application, and as teachers, we can't always track that, right? Like if I'm teaching the difference between phonemic awareness and and phonological awareness, which I feel like, again, I say I've said this frequently, and I'm still not alone that it takes my students a while to get to the point where they're really at ease with that, but if we use a tool like AI that is designed to give them more opportunities to respond to something that maybe they're questioning the student and giving them feedback in that way too. I don't have time to track that for each of my students, but if there was a some sort of a resource that could help do that, I think that would be one good way to use it.
Donell Pons:
Yeah, we've talked about the lack of time too that you have, Stacy, for teaching such large concepts. Wouldn't it be nice to find a way to kind of accelerate some of the learning, right?
Stacy Hurst:
Yeah, for sure. In
Donell Pons:
that regard, Lindsay, what do you think? Ever have an application in a first grade setting, even if we're intentional?
Lindsay Kemeny:
Well, I think when like you're you're kind of you're asking like when in the day maybe could I foresee and does it have
Donell Pons:
a real advantage if we were intentional with it? Could it really yeah first grade setting? Yeah, I can
Lindsay Kemeny:
see I can see it being helpful during our small group, what I call small group and centers, when I am meeting with a group and I have my the rest of the students doing something, I can see you know them could be helpful if it was accurate and walking them through. Now there's different products. Out there, I think right now, which I hear a lot of bad reviews. So we just need we need refinements there. But where they are reading aloud, and then the you know the the chatbot or the AI or whatever the program recognizes what they're saying. So that's where it's usually a problem. Is it doesn't accurately understand or recognize what they're saying, but being able to give them feedback right there in their reading as they're reading aloud, and maybe walk them through the step of sounding out a word correctly if they missed it. I think that could be helpful. Maybe helpful walking them through asking questions about a passage afterwards. You know, maybe finding the gist of it and walking them through that, it would just it would just have to be accurate because otherwise, like right now, I feel like what's out there right now is just tends to be frustrating for students and for teachers. But you know, that's that's like the opportunity cost of me working with a small group as students are working independently. So I think that potentially is where, like, as a teacher, I would want some more help because I can't help them right now. I'm I'm with another group,
Donell Pons:
right? Yeah, that's interesting. And you brought up the important point of accuracy, so that it doesn't lead to further frustration for a student, right? Making sure we don't rush to market with things that maybe aren't quite ready.
Stacy Hurst:
Pass information, yeah, like not check thoroughly, yeah. I also think again that human aspect, the key I feel like for educators is to stay in control of it. Like meaning it. You use the word intentional, Donell, which I think is perfect, and you said it too, Lindsay. Like intentionally, you're not replacing you during small group time. What they could potentially be doing with something that is AI driven would be better than not having a teacher involved with them right at that moment. But they're still not the teacher, and I and I agree from what I've seen. And I'm by no means an expert on what's out there with AI, but I have I have examined some tools that can fairly accurately. I don't even say that with a lot of confidence. Capture what a student says and compare what they should have said to what they did say. But the thing that I haven't seen anybody be able to capture yet is that human aspect of it? Are they saying that with frustration? Is their voice, you know, exhibiting that prosody we look for? How do we measure that? Right. I think there's still such limitations that we cannot rely on it wholly. Like the teacher has to be the one making those judgment calls. I think,
Lindsay Kemeny:
and I just think we have to be so cautious with whatever is happening and whatever is being implemented. I think we need to wait for a lot of checks on that and parameters, you know, barriers around the AI because I think it could go south really quickly. And sometimes you just you you wouldn't even think that something would come up, and then some suddenly, you know. So you just want to be careful if that student is having conversations with the AI and it's answering back, you know that it's you know that that that there are some limitations there because even just you know I remember when it was one of those devices like an echo.or something that you know was very new, and there was a teacher at our school who who got one in the classroom, and it was very quick that my my child was in that classroom. It was pretty quick that I got a call from the school, and someone had asked it a question, and the response was sexual in nature and was sharing some things that shouldn't have been shared in like a fourth grade classroom, and we got a call about it as parents letting us know. And I'm like, that's the kind of thing like you just we we shouldn't be jumping into these things, but we should be more cautious.
Stacy Hurst:
And we are not there yet. That is a very real concern. Yeah,
Donell Pons:
yeah. I'm glad you brought up the point about the personal nature of AI and how it's different from other technology because that was brought up in some of the the reporting in the Stanford report was to think about that with AI in particular. Is is it okay to encourage students to use something that they may be then having a relationship develop that we hadn't thought about before in a way we haven't thought about before. It's really interesting. And another thing, while we're on this juncture, is that a lot of the AI they're discussing here and the use of AI is English based, and so they were also talking about the sensitivity to students who speak other languages and making sure that does it provide something of at least equal quality in that student's language to make sure because that isn't always the case.
Stacy Hurst:
Or even dialectation.
Donell Pons:
Yes, even dialect. Yeah, has the potential that it's just making sure that you it's as great as you think it is right before introduction to students. Okay, so let's talk about a positive as we kind of. Are winding down our conversation here. One of the positive things was that AI may be particularly valuable for a certain group, and that would be teachers and administrators. And one of the things I said about teachers is that teachers using ChatGPT with a specific guide spend approximately 25 fewer minutes per week on lesson preparation without detectable differences in lesson quality, according to blind expert ratings. So, what do we think about that as a tool for teachers?
Stacy Hurst:
Well, right away, this is my questioning brain with with no effect on lesson quality, but was there an effect on student outcomes? I guess that would be the maybe that's what we can assume or infer by lesson quality.
Donell Pons:
Well, it says here supporting less experienced educators in the tutor copilot study. Students of less experienced and lower rated tutors experienced larger improvements in topic mastery. The report identifies these findings as promising, while noting that more research is needed to understand the conditions under which AI supports educators effectively.
Stacy Hurst:
Yeah, and I think Chris Agnew spoke to that at Big Sky as well. Yeah, that it did have a benefit. See, it all goes back to teacher knowledge, and so it might have had a benefit for teachers who knew less, but then also did that help them to learn more, or were they just over relying on it eventually? I'm a teacher educator, so I, you know, those are the questions I would want to know. You always want to improve what teachers know.
Donell Pons:
Yeah. What do you think, Stacy or Lindsay?
Lindsay Kemeny:
Ah, like I'm all about saving teachers' time. This one worries me a little bit with lesson planning, just because where is AI getting the information, and is it the best kind of research-based approach that it's sharing? Because I don't know that it knows that, so that would worry me a little bit. I mean, you said it didn't affect quality, but really, how did they how did they rate that? You know, but that being said, and it's not fair if a teacher is having to use it for for lesson planning. The first problem I think there is maybe they don't have a program, and so that would be a district problem where they need to fund their literacy instruction in their district. So I don't know. I'd probably want more specifics on how exactly it's being used. I have used AI to help me write difficult parent emails. So you you're like, I want a kind. You know, how can I best explain? Like, I don't want to say your child is bossy or something. You know what I mean? Something measured. Yeah, it can be kind of helpful. I think in that, you know, I I don't know, you know, there's there's probably that's what how I've used it mostly. I'm sure there's other like great ways, but again, it's just like being cautious. But I'm all about teachers saving time. So like we have 1000 things to do, and it's amazing. It's amazing. Everyone is still a teacher because of everything we need to do. We love it, but there's a lot. It's a lot. But
Stacy Hurst:
you know that is how I use it because I'm prone to overthinking things, and I can get stuck in these thought loops. And so I use AI to get me unstuck. Now it doesn't mean that I just take for what what it says at face value, but
Lindsay Kemeny:
yeah,
Stacy Hurst:
I'll say you know I have this, this, and this, and it all needs to be done today. What's the priority? And I already have ideas because I've been overthinking it for an hour, and then it will say, well, you can look at it this way, this way, this way, and then it just helps my brain go, okay, I can, you know, I can make a choice. Go, and that's
Lindsay Kemeny:
a good point, Stacy, because even when I've used it, like help me with this email, I always I'll be like, oh yes, that's that's a better way. But I always go and take it and like adjust it a little bit for the email. You know, I don't use it, and it's just the same thing with teachers using it for anything. Like if they're using it for a lesson plan, to make sure to go back and very carefully analyze it, don't just rely on it. So I see I've seen teachers online sharing like AI written lesson plans, and then they're selling them, and they're or just like visuals. And I'm like, okay, like I've looked at this for 10 seconds, and I already see like two errors right there that are pretty, you know. So and I know the AI just spit that out, and then the teacher didn't go back and like carefully, you know, review it. So yeah. Anyway, just a thing to like, like we said, it's not always accurate, so you you have to review it.
Stacy Hurst:
And what it spits out is only as good as the prompt. Yeah,
Donell Pons:
and and remember the the the research we were referring to did say teachers using Chat GPT with a specific guide, so they were
Lindsay Kemeny:
so they were set guidelines and yeah yeah
Donell Pons:
spent approximately. So this was very structured. In other words, you weren't just just out on Chat GPT doing something. So again, we don't have information for that, do we? Yeah,
Lindsay Kemeny:
yeah.
Donell Pons:
I thought it was interesting that for scheduling, and Chris Agnew brought this up at the Big Sky Literacy Conference, who's involved with the Stanford AI research, brought up the fact that for scheduling in a school, so for this is the administration piece, it could be rather helpful in lifting that load because that takes a lot of time.
Stacy Hurst:
Yeah, I think that's great. I mean, I always
Lindsay Kemeny:
think about that. Like with our poor vice principal, I'm just always like, I can't imagine doing this schedule because you got to figure out, you know, when are they going to preps? When is recess? When is this? Is this? So I think that is like awesome.
Stacy Hurst:
Yeah, as someone who's been in charge of scheduling those interventions across a school and across grade levels, and that would have been amazing. Yeah, would have been amazing tool to have. And
Donell Pons:
then, as always, we want to caution that if anyone is using AI in a school setting, that you you make sure if you're not using a specific plan that you have checked with your district or with your state office to make sure you're not sharing information that should not be shared, and you can receive guidelines. So along the way, we're going to keep needing to check on this to make sure. And eventually, I'm assuming we'll have that locked down. But for now, it's kind of the wild west. You need to make sure it's not always what you think, right? It's a good point.
Lindsay Kemeny:
Careful putting in data and things like that,
Donell Pons:
yeah, things like that. So I think this has been really interesting. Just to recap, because we talked about a lot in our conversation here, is AI can improve immediate performance without guaranteeing any durable learning. That was one of our first points that we put out there. The design of the AI tool and the cognitive work students perform really matters, and the evidence base is still too limited to make broad claims about AI's effects in the U.S. particularly in K through 12 education, and particularly for literacy as it comes. So stay tuned. There will be much more to come. But I loved having this conversation. Thank you, you guys. Very interesting. Yeah,
Stacy Hurst:
thank you. There's still more to think about, I'm sure. And Donell, you've given us a lot to think about in a framework that is manageable, because there's a lot when it comes to AI, and we have been narrowly focusing on AI in education and for learning. But I think there are broader implications for society. Like you said, we could go on about this for a long time. But thank you, thank you for guiding this conversation, and thank you to those of you who've joined us for this episode. And we hope you join us for the next episode of Literacy Talks.
Narrator:
Thanks for joining us today. Literacy Talks comes to you from Reading Horizons, where literacy momentum begins. Visit readinghorizons.com/literacytalks to access episodes and resources to support your journey in the science of reading.
