Artificial intelligence is increasingly common in classroom instruction. The growing adoption by teachers and students has opened new realms of possibility, and also begged many questions about proper use.
In a recent public lecture, VAI Learning Specialist Ben Talsma offered keen insight into the benefits of AI in the classroom and a nuanced take on the pitfalls.
Watch the lecture below:
Video transcript
Note: The following transcript has been edited for readability. Click a timestamp to jump to that part of the video.
Maranda:
Hello friends. We are so glad you have joined us today. Good afternoon. Welcome to Van Andel Institute’s Public Lecture Series. I’m Maranda and it is exciting to have you here joining us in person and virtually. I’ve had the opportunity to speak with our presenter today and you are in for an exciting time. We will be talking about a hot tech topic: AI, artificial intelligence. I’m sure everyone here has heard about this latest advancement, especially when it comes to AI and how we all are using it every day. Whether we realize it or not, it quite literally is everywhere. It’s in your phone, it’s in your office, it’s in your smart home. And, of course, it is in our classrooms.
Schools across the world are using AI to boost efficiency and effectiveness, adapting educational methods to incorporate various new technologies. And today you will hear about these and you’ll also hear how VAI is fostering ethical AI use, which is a big topic for so many, how we’re doing that in educational settings to help ensure that students have the best and the safest learning tools. AI is quite complicated and for many of us it’s probably kind of scary. We can do so many different kinds of things.
Thankfully, we have a member of VAI’s Institute for Education to walk us through. Ben Talsma is a passionate, innovative educator with more than 20 years of experience in the field. He has taught students from kindergarten through middle school and now supports teachers in schools at VAI Education. He has been an early adapter of AI in the classroom and his insights on the topic have been featured in the Detroit News, Chalkbeat Magazine, The Week, e-School News, and on the podcast Overthrowing Education. After the presentation, we’ll have time for questions and answers. Those of you joining us online, you can also put your questions in the chat. Let’s get right into the topic and the discussion. Join me in welcoming Ben.
Ben Talsma:
All right. I believe the publicized title of this presentation today was “How K–12 educators are navigating AI in schools.” It’s different from the one that I have on the board here. I did that on purpose to change things up here today, to put it in the form of questions. How will AI change education? I did that because I’m an educator, because I like promoting a little bit of curiosity and because I want you to think about how you might answer that question, what sort of preconceived ideas you have, what sort of thoughts you have about how you would answer that. So, while you are being a little curious and wondering about the future of AI in education, I am curious what you come up with. So, go ahead, raise your hand if you know the answer to that question. How will AI change education? <Laughter>
I hear some laughter. I see no hands going up right now and that is okay, because you are in exactly the same boat as all of us in education right now. When it comes to making long-term prognostications about how this technology is going to be impactful, it is still a complete mystery. So, I am not gonna be speculative in the long term. We’re not gonna talk about whether or not AI will take over the teaching profession and render your teachers obsolete. We are not gonna explore whether AI will eventually turn the world into one giant paperclip. It’s actually something people are concerned about right now. We’re not gonna think about the rise of machines. Or, on the plus side, we’re not gonna think about whether we’ll become hybrid creatures of fewer lights who are gonna travel the cosmos on electric rays, discovering all of its mysteries.
We’re just gonna look at what’s happening right now in the classroom. So, I am a teacher professional developer. At least a lot of the time, I am. And so I’m able to get into classrooms and talk with teachers, to talk with schools about how they’re putting these technologies, this generative AI technology, to use for them. Some of the things they do are things that they try, and they flop, and they tell me about that. Some of the things they do are things that they try and they love. And so I’m privileged to be able to serve as a kind of filter to hear about what’s working for schools and then to disseminate that — to spread back to the other schools with which we work. To share out the best practice about how these tools are being used to move the needle in education. So, I am excited to be sharing those ideas with you today. That is an AI rendering of myself. I think it does a nice job of capturing my level of excitement about how thrilled I am to share these ideas with you. Those other two people are AI renders of you. I think they got you spot on.
As we do this, my hope is that you’re able to get a feel for what it’s like inside classrooms right now — to get a feel for what it looks like to see these technologies being used in K–12 education. And in order to do that, we’re gonna look at it through three stories, three lenses. So, first, we are going to tie my daughter to a tiger. Second, we’ll be looking at the tragedy of the shoelace. And third, I’ll tell you the tale of the baboon and the weasel. And each of those is gonna show us something a little bit different about how schools are making use of AI, and how AI is impacting education. Ready to rock? Excellent. Alright, so to start off, I’m gonna show you one of my favorite pictures of one of my very favorite people in the world. This is my daughter, Violet.
She is a dancer and a discus thrower and a swimmer who’s always been comfortable in the water. And I think this captures a lot of that in that one joyful pose. Now, somehow or other, she is no longer eight days old. She is no longer eight years old. She is now 18 years old, which means that she recently did this. And it means that in just a few hours, in fact, as soon as we’re done here, we’re gonna put her on one of these and she is going to fly all the way across the country to western Washington to begin her life as a college student. It’s a big change. And I gotta tell you Violet, who’s in the audience today, we are so proud of you. I’m a proud father — proud of all that you have become. I’m feeling a lot of pride. I’m also feeling a lot of excitement. <Laughter>
Like the graphics there. A lot of excitement about all the new adventures. And I am so terrified. <Laughter> Not because there’s any specific cause to be terrified, but just because — can see me, and this is me again by the way, different version, see me in Caveman gear, right? Because there’s something deep within us that’s really apprehensive about shifts, about surprises, about change. This goes back to the days of yore, maybe even before the days of yore, when we were lurking around on savannah or in jungle, and we needed to be attentive to change because if there was a sudden change, there’s a pretty good chance it meant that you were going to become delicious. <Laughter> Now, there were good changes. Perhaps there was a new source of berries around that, you know, form. But for the most part, the risk outweighed the reward. So that’s how I’m tying my daughter to a tiger. Both of the situations — getting on the plane and flying across the country, confronting a tiger — involve change, and activate some of that apprehension that we have. We are programmed to fear change. That is me fearing change, right? <Laughter>
That’s the same sensation that teachers felt when these AI tools kind of burst out into the public consciousness. When AI, kind of, became something that they were aware of and they started to learn about its capabilities, the first thing that happened was they got worried. So, we are going to look at some of the top teacher terrors. We’re gonna look at — sort of issues that they’re wrestling with and some of the fears that they’re trying to mitigate. Let’s get going.
The very first one, the first fear that teachers came up with, was the fear of plagiarism. They showed their nice little essay prompt that they’ve been giving for 14 years to ChatGPT, and in about 30 seconds, it came up with something that just hit all of those criteria extremely well and it made them really worried. How can they be sure that the work that their students are turning in is actually the work of their students’ minds?
And of course, there’s constantly this struggle between whether students are gonna get around the teachers trying to catch them plagiarizing. Kids can be sneaky, right? They can do things like adding a couple grammatical mistakes just to make it look a little bit more authentic, or using the vocabulary of a seventh grader, or writing mainly using simple sentences. These are things that ChatGPT can do if you ask it — that all of these large language models can do if you ask it. And so the first thing that teachers thought was “plagiarism alert.” Let’s try it out a second. We’re gonna do a little ChatGPT in-crowd game. If we can switch over to my computer right now, see if it pops up there. There it is. Teachers try and get creative. They try and work around this to try and get an interesting challenge that some computer’s just not gonna be able to solve. So, we’re gonna do this one in poem form. Somebody shout out a type of poetry.
Audience member:
Haiku.
Ben Talsma:
Ah, a haiku. And give me some academic content that you might be teaching in your class.
Audience member:
Integrates.
Ben Talsma:
Integration in calculus. That is a challenge, right? And whatever challenge teachers can come up with, this can probably figure out an answer. Really tough in a text-based world to work around the ability of these tools to produce really good stuff. If you want it to do a sonnet, it can put it in sonnet form. Essentially anything text-based, it does an incredibly good job. So, teachers are really struggling with what to do about this. If there’s something that can answer any textual prompt that you give, what in the world are you gonna do? Let’s switch back over to the PowerPoint presentation.
So here, well look, let’s explore some of the other concerns. First, plagiarism is one of the top concerns. Related to that is the question of brain decay. If students aren’t using a particular skill, they’re gonna lose it. It always makes me think of the cabbies in London who have to be able to navigate this famously intricate city. They’ve gotta know all these twists and turns. And when they study their brains, they discover, they find that the visual spatial sections of the brain are much bigger and more developed than the typical person’s. But, those big brains in a world that has GPS, the one thing they discovered is, to the extent that cabbies make use of that GPS technology, the less pronounced that development is. In a very real way. If we don’t use it, we will lose it. And so that is a valid concern that teachers had as well.
Teachers’ third concern, the third thing that popped up, was the question of misinformation. Because AI tools, large language models do have a tendency to hallucinate, to come up with things that didn’t actually happen or to connect quotes that were never said. So, misinformation was something — they don’t want kids learning wrong stuff from AI. Tardigrades never invaded the White House — just didn’t happen. <Laughter>
And then last, whenever you’re thinking about tech use with students, there are concerns about things like data privacy or student misuse of this technology. And so those are four of the big concerns — not all of the concerns, but four of the big concerns that pop up in teachers’ minds.
Let’s look at what they’re doing, because certainly their tiger terror cortex has been activated, right? That fear of the unknown, that fear of change was activated. But teachers are a resourceful lot and so they are developing toolkits that they can use to fight back against some of those negative repercussions. When it comes to plagiarism and brain decay, we’re doing a lot more in the classroom right now. How many of you wrote in any Blue Books at some point in time? It’s coming back. Because if students are handwriting stuff in class, we can be sure that what they’re thinking is what you are seeing in that book. We’re doing more dialogic assessment where students have to share orally what they know. They respond to some questions from a group, from the class or from the teacher.
Performance tasks are something that more and more classrooms are interested in. Performance tasks are, I think, a great analogy is a bicycle. If you give somebody a test about how to ride a bicycle, they could answer all of those questions correctly and still not be able to do it. The test of whether you can ride a bike is getting on one and riding, and performance tasks are the academic equivalent of that. They’re things like, instead of answering questions about a scientific method, can you construct and carry out a good scientific investigation? Instead of knowing about building a budget, can you actually use one in a market day kind of environment, right? So, taking their learning and making it real is a way to counteract the concerns about plagiarism that teachers have.
Here’s a life-hacking kind of one. I like this one. It’s clever. This is called sneaky white text, as you can see by the text becoming sneaky and white over there on the right side. <Laughter> So you’ve got your prompt, this prompt that you’re worried about kids just copying and pasting and tossing into a large language model and getting an answer. Then you add something that’s a little off topic. For example, you might ask your Jane Austen essay to include references — to have at least two references to baboons. You take that text, you make it white, it disappears. And then anyone who copies and pastes it and just sends in whatever the AI system produces, writes you a very nice essay on pride and baboons. <Laughter>
If you don’t want it to be that obvious, you can make that connection a little bit less ridiculous, right? George Elliot came long after Jane Austen — couldn’t possibly be somebody who Jane Austen was influenced by. So it’s a little less obvious, but it’s a clever kind of hack-y way to show kids that yeah, you’re, you’re looking out for this, you’re with it, you’re paying attention.
But, it’s not all about catching the kids. Ideally, we want kids to be able to develop their own sense of when and when not to use these tools — how and how not to use these tools. So, this is one of our free resources that we share with schools all around the country. It is a collection of different scenarios about whether AI should be used or not — or scenarios in which AI is used. And then students need to evaluate how right it is from extremely, perfectly fine, maybe using spell check to help you catch spelling errors before you turn it in, to completely unethical, having AI do your entire essay and then passing it off as your own. And students get to discuss and debate why they might move something in one direction or another. They get to work together as a class to come up with characteristics or principles that help guide ethical AI use in the class.
They can come up with a policy there, either as a class or maybe as a school, based on what they think the purpose of AI can and should be for that topic. It’s gonna be very different, you know, in different classrooms and different subjects, but it helps teachers and helps students to develop their own moral understandings of when and when not to use AI — how and how not to use AI.
When it comes to misinformation, we’re seeing lots of teachers who are sharing AI-produced content so that students can fact-check it, being intentional about getting students an opportunity to fact-check what AI is coming up with. Or a little wrinkle on that, asking students to stump AI or to bait it into giving a wrong answer. Here you say, “Can you give me a prompt where AI will answer incorrectly?” without saying, give me something ridiculous or give me something false. Can you find the places where AI really struggles? Great metacognitive task. A task all of us are working on now when we use these tools because we want to know what sorts of things these do well and where do they really stumble.
So, asking students to do that same sort of thing is a really interesting intellectual exercise that exposes them to how these things work. Plus, whatever it comes up with, then they get fact-checked. Does that make sense? Wonderful. I also love playing “beat ChatGPT.” We’re seeing a lot of teachers do this, where they share some AI-generated content and then they ask students to improve upon it.
I love this because this is super authentic to how AI is being used and how it’s going to be used for the rest of their life. For the rest of their life, they’re gonna be challenged. Can you do something better than the machine? Can you add value to this? And, so, giving students a context in which to practice that is extremely valuable. Plus, it’s a great academic exercise. Study this content, find where it’s wrong, find where it could be improved. It’s a real great challenge for kids.
I also love how this shifts the mindset. Instead of thinking of ChatGPT as a search engine — it’s just gonna give you the answers — you become a much better thinker when you think of it as a collaborator. ChatGPT is one of the smartest, if not the smartest, person or the equivalent of the smartest person in the world. And you can talk with that person whenever you want, and you can learn from them, and you can bounce ideas around with them. And it won’t always be perfect, right? You can still improve on it, but you have the opportunity to engage in a productive, collaborative discussion with a tool like ChatGPT. When it comes to data privacy and misuse, there are a couple of really great tools that are coming out.
Lots of schools use MagicSchool. My personal favorite is SchoolAI. These are tools which allow you to customize what your kids can and can’t do when they’re interacting with artificial intelligence. So you can say, we are only today talking about beans. And if they try and go off topic, you’ve got a little dashboard that flags situations and says, Violet is not talking about beans anymore. Maybe you should go over by her and get her back on track. So, it allows teachers to expose students to some of these AI tools in a way that has good guardrails, that provides good guidance.
All right, it is recall time. Not recall time in the disastrous Ford Pinto sense, but recall time in the sense of remembering. Hold on, you’re probably sitting relatively close to somebody. If you’re not, you can just think to yourself. But try and identify, try and remember what those four big concerns were and what teachers are doing about them. Just two minutes to discuss or reflect together. Go ahead. Can you remember all four?
Audience:
[Audience discusses]
Ben Talsma:
All right, show me on your fingers. How many could you remember? Cory thinks he remembered all four. I don’t know if I believe him. <Laughter>
Repeated recall, remembering information, is a great way to help solidify it in our minds and our memories. One thing that I want to demonstrate throughout this whole discussion is that good human cognition hasn’t changed. When teachers use these tools really well, they’re doing it in a way that facilitates really deep human cognition. So, tying my daughter to a tiger is all about risk.
Now, I wish to tell you the tale of the tragedy of the shoelace. This is a story about a teacher who is younger than I am right now, who is lighter than I am right now, who is almost exactly as good looking as I am right now — I think. He was rushing off one day to run some copies and he had to get those copies and get back to his classroom in time to set up the learning activity that was gonna take place.
He had enough time to do that., but then he looked down and he discovered something which shook him to his very core, and that was his shoe was untied. And he realized that there was no way in the world that he was gonna bend all the way down, tie his shoe, stand all the way back up and still have time to do all the things he needed to do. And, so, he had to go with his shoe untied for that time. If you haven’t been in the classroom, I see some teachers in the class — in the room — and they’re nodding their heads. If you haven’t been in the classroom, it’s tough to understand how busy some teachers get sometimes. In fact, lots of teachers get lots of times. There are lots of situations where you don’t have time to stop and tie your shoe.
So, real big issue in education. In fact, if I was gonna make one big change, it would be to create the gift of time for teachers. And with AI, I can kind of do that. One thing that teachers are using AI for in a really positive sense is saving themselves time. So, if you want to come up with a lesson plan — say you are working on a lesson plan for the rock cycle, you can go to ChatGPT or another large language model and say, “Hey, could you make me a lesson about the rock cycle?” And in 30 seconds, there it is. You can change it, twist it, make it your own. If there’s something you don’t like about it, you just ask for it to change. If you don’t like it at all, you can say, “Gimme five different ideas” and pick from the one you like best.
But, instead of having to go through that whole process, you can get that draft and it’s much faster and more effective to respond to that draft than have to generate it yourself. And what that does — is it frees teachers up — now they’ve got a little time. We can be a little more creative as a professional. So, if you’ve got some students who are big fans of superstar actor Dwayne “The Rock” Johnson, you can say, “Hey, build me a lesson plan that ties the rock cycle to Dwayne “The Rock” Johnson.” And in 30 seconds, just like that, you’ve got a really nice lesson plan. So you can tap into that creativity, you can connect with your students’ interests. You can do things in a little different, more novel, more interesting way.
Another way we’re saving time is through differentiation. This is a picture of me looking guilty because, in the classroom, I always wish that I could have done more differentiation. It takes a long time to take an assignment and to say, “Let me make a little easier version to scaffold students who are struggling. Let me make a little harder version to challenge those students who need that challenge”.
Now, we can do that in an instant. You can take the same content, and you can have it written at a lower grade level, and on grade level, and a higher grade level. So students are still learning the same thing, just with different texts which are appropriately challenging for them. You don’t even have to have it in text form. You can now take a picture of that assignment, upload that into a large language model like ChatGPT, and it will understand what’s in there and be able to make those adjustments — at least a draft of those adjustments — for you to more directly support and meet the needs of your learners.
That’s differentiation for readiness, for student achievement. Teachers are also differentiating based on interest. That’s a picture of a baboon ’cause I’m interested in baboons. I think it’s a great, fun word to say. If you ever need to cheer up a little. Just say, “baboon.” <Laughter> It works.
So, if I was in the classroom right now and we had this reading assignment where we had to read a passage and answer some questions, my teacher could take that and, just in a flash, make it the same basic idea, the same challenge level of questions, but all about baboons. It’s a wonderful way for students to get tied in, to get pulled into their assignments. When teachers differentiate based on interest. And it works for anything, it doesn’t have to be baboons. It can be kazoos. It can be balloons. The possibilities are limitless. It also can provide drafts of feedback, right?
If you’re an English teacher, you’ve got a hundred essays to respond to, it can give you that rough draft. You always as a teacher want to have that executive authority over it, but it can give you a draft to respond to so you don’t have to do all of that laborious — we call it “thunking,” right — the thinking part that doesn’t really involve higher-level thinking. It can function as a teaching assistant, right? It can do that grunt work. You can function as the executive, as an educator. You can kind of change it up, make sure that it matches your expectations, your classroom, but it’s taking a lot of that work off your plate so that you can focus on something else. Building relationships, designing new interesting experiences, right? Or maybe just relaxing ’cause teachers need to do that as well. So, things like drafting newsletters, emails, or sub plans — analyzing data. It can do a lot of these things to save teachers time and give them the opportunity to recharge and refresh in important ways.
So, tying my daughter to a tiger was all about mitigating risk. The tragedy of the shoelace was all about saving time. And last, but certainly not least, we’ll explore the tale of the baboon and the weasel. Again, I’d like to hearken back to my days in the classroom because I was a pretty good teacher in a lot of ways. I was a really good teacher in a lot of ways, but I couldn’t draw. Some original Talsmas on the board right here. <Laughter> If I had an idea in my head, I just couldn’t get that idea out there on paper or on the whiteboard, so I was excited when I heard that these tools could take an idea and turn it into an actual, real live picture.
Very first interaction I ever had with an image generator was, I went to the tool called DALL-E — way back then, DALL-E 2 — and I said, “Create a surrealist masterpiece featuring a baboon, conducting a duel with a weasel, with a stick.” And this is what it came up with. Twenty seconds. And all of a sudden, it blew my mind. I thought, oh my goodness, those two ideas that existed in my head, now I can bring those to life. I did it exactly one year later, and this is what I came up with, so it got better fast. And now, I had a superpower.
So, one way to think about the power of AI is to think about it as a tireless teaching assistant. Another way to think about it is that it can instill in teachers superpowers they never had before. So, if you could not draw, now you can create pictures. If you could not code, you can now write code. You can make a video game using these systems. If you don’t speak Spanish, now you can communicate with somebody in Spanish. It can instill in our teachers’ superpowers.
So we now, if teachers take good advantage of this, have superpower teachers running around out there in the world, which is a good thing. It’s a good thing. They will use their power for good and not for ail. It’s a whole new world where these kind of things, having lots of creative ideas, being able to choose the best ones and having fun, bringing them to life is increasingly important and increasingly possible. So, teachers are using these AI tools to teach deeper. One really powerful way they’re doing this is by helping students learn from examples. We learn really well when we learn from examples, when we see different, different characteristics out there in the wild and we pick and choose what’s gonna be important to this category we’re constructing.
This is how you all learned what a dog was, right? You saw lots of dogs, you talked about all those dogs. Sometimes people said “No, that’s not a dog, that’s a cat,” right? And eventually you became one of the world’s foremost experts on dog identification. You can tell a dog from a cat from a mule 10 times out of 10. We can do that same sort of thing in all sorts of different domains now by helping our students to learn from examples, because AI, because ChatGPT can generate them so quickly and easily.
So, one thing that lots of teachers are doing right now is saying, “All right, I’ve got a concept. ‘AI, could you please get four versions of it? Good, medium, bad and terrible.’” Then sharing those examples with students and allowing them to rate them or to rank them, to characterize which ones are better and which ones are worse, to think about what characteristics make one better or worse, to really think deeply about the content that you’re exploring. Then, they can have discussions about what they ranked and why. And we find that when students have done that initial ranking or rating, they have these great conversations about why they gave one of these pieces of descriptive writing four stars and another three. And they can work together to identify, construct their understanding of what makes a good example compared to a bad example. Teachers are sharing lots of B+ exemplars — examples that are good but not great, again, to challenge students to find the good in it, but also to view this as something that they can improve on, right? So this is a great way for students, for classrooms to construct an understanding of the characteristics of what make any kind of work great.
And when we can construct our own understandings, we learn that thing more deeply. We apply it more broadly. We remember it more durably. I love this one. You can have students have a conversation with anyone throughout history. So if you want your students to chat with Frida Kahlo or Vasco da Gama, you can have AI role-play that character so that your students can ask questions of them and follow their curiosity to learn how that person actually lived or actually would have responded.
And it doesn’t just have to be famous people. You can do this to help students see what it was like for the, the underrepresented people, right? The peasants back in the days of the Wars of the Roses. You can have them interact with literary characters. You can talk about Templeton’s perspective in “Charlotte’s Web” with Templeton. It doesn’t even have to be a person. You can talk with a leaf about what it’s like to photosynthesize. Now, these tools like SchoolAI allow you to keep that conversation focused, allow you to build what sort of conversation the students should have at what grade level. It allows you to identify misconceptions your students have so that the AI system makes sure to address those. So, if you’ve got lots of students who feel like leaves have little sand feet, you can make sure that the AI system disabuses them of that notion.
You can allow students to follow their curiosity, right, to ask the questions that they’re curious about or to find things out that they are wondering or that they misunderstand. So you can allow students to personalize their learning journey by using some of these tools that help to guide how your students — or what your students — are exploring. I love that these allow us a chance to get playful.
One of the coolest ways that we’re seeing teachers put these tools to use is to come up with academic versions of common games. How many people have played Apples to Apples? Most of you, great. You can build a version of Apples to Apples that’s all about habit-habitats and ecosystems. In two minutes, you can have this game where students are having all these discussions and debates about which adjectives best describe which category cards and opening up these rich conversations where students feel like they’re having fun, but they’re learning in a deep sort of way.
Here’s a version, again, of Taboo that’s specific to particular academic content. So you can take games and come up with versions that really help your kids dive into the things that you are learning. Scattergories is a great one for this. So, it’s helping our teachers be more fun in a way that doesn’t compromise depth of learning at all. In fact, it enhances it. This is one that’s being used at the beginning of school all over the country right now, coming up with name tags that represent your students directly. So, I’m gonna actually call my daughter up here for this. Violet, would you come up? Can we switch back over to my screen, please?
So, say that Violet was a student in my class, and I wanted to create a name tag that reminded me of the important things about her, some things that she likes, allow me to connect with her. Violet, what’s your favorite animal?
Violet:
A blue giraffe.
Ben Talsma:
Blue giraffe. What’s your favorite subject in school?
Violet:
Psychology.
Ben Talsma:
Oh, psychology. What’s a favorite hobby of yours?
Violet:
Swimming.
Ben Talsma:
And where’s a favorite place of yours?
Violet:
Mountains.
Ben Talsma:
Swimming in the mountains.
Now, once I’ve got this name tag, I can put it in onto a name tag, little placard, and put it on Violet’s desk. And then, every time I look over there, I can see some of the things that she’s interested in. Some of the things that she likes, that she cares about, allows me to form connection. Hey, look at that. So that is my Violet name tag. I need more psychology. I would fix it to add more psychology in there. But you can get that picture, right? A fun way for you to understand and remember things about your students. They can look at those name tags and they can see where they have things in common. Give Violet a nice round of applause. A way for teachers to connect with their students.
Alright, we’re approaching the end here. What do you remember? Go ahead and turn and talk with somebody near you about how teachers are using AI to save time and to teach deeper. What can you remember? Go ahead, one minute.
Audience:
<Audience discusses>
Ben Talsma:
Alright.
Ben Talsma:
One thing I love about this is it’s a lot easier to remember stuff when you’ve got a picture. It’s called the picture superiority effect. And, so, teachers who are using AI image generators to create those pictures are really helping students to remember things as they go. Alright, we’ll bring it into a close. What are the big ideas here? Again, I hope that you took away that the ways in which teachers are using this technology are all about deep human cognition. The things we’ve always wanted to facilitate in class. We want students to be curious, to be creative, to be critical thinkers. We want them to collaborate well. We want them to view learning as an interesting adventure they’re exploring. We want them to have fun while they’re doing it. And we want them to use what they learned to make the world a better place.
And if we use this technology to accomplish these things, it will be a good technology. Every technology is what we make of it. And if we can help create classrooms where tech — this technology is used for good, it will be a force for good in the world, because that’s what we’re all about right here in this building. We have people who are using their skills and their understandings to have a positive impact. They run around this building every day using the latest and greatest and most appropriate technology. And our students, the students who are in our teachers’ classrooms, are in the same position. They’re gonna grow up, they’re gonna be the next generation of people who use technology to make new discoveries and to have a positive impact. Teaching is hard. Professional development is hard. It has a lot of inertia. But these tools — I’ve seen them have those light bulb moments — have those mind-blowing moments where teachers realize they now can do things in a way that they couldn’t before.
So I hope you got a good look at how we’re mitigating risk saving time and teaching deeper. This is a time of incredible change, and that change can be scary, but if we learn to embrace it and embrace the positive opportunities, then it can be a wonderful time full of exploration and discovery — little metaphor there. Thank you for coming. We’re gonna have a question and answer period, and then we’ll have that tour down at the building. You can ask me about anything that we do. If there’s a school who you love and you want to put us in touch, just let me know. I’ve got a card I can pass out. But I think now is a great time for Q&A.
Maranda:
Okay, let’s have a seat. Phenomenal. <applause> First of all, I want thank you for taking something that is scary to a lot of us and suddenly getting me very excited about it. I’ve seen a lot of presentations on AI, and there’s always that fear factor. You’ve removed the fear and said, “Let’s look at the possibilities for good.” So thank you.
Ben Talsma:
Wonderful. Glad that, glad that works.
Maranda:
So, for those of you in the audience, if you have a question, we have microphones that will be passed to you. We’d like you speaking to the microphone to ask a question. If you are joining us virtually, go ahead, use that chat for, or ChatGPT, whatever you choose. <Laughter> Go ahead and post it and we’ll try to get you as many of those as we can.
Ben Talsma:
Someone out there should ask ChatGPT what question they should ask about. <Laughter>
Maranda:
You really ought to. I — I also want say, I love the connection between Ben and Violet, and isn’t that a beautiful send-off, man? You should be so proud of your daughter, and you should be so proud of your dad.
Okay, friends, what questions do we have? I’ll get it started. What are you afraid of when it comes to ChatGPT? Do you have those fears that we talked about or are you over it?
Ben Talsma:
So, I think the most interesting fear I have right now is the, the brain decay question, right? Whenever we get a new technology that changes the way in which our brains operate, right? The GPS example we shared is one, but they think that when we domesticated dogs, our brains shrank a little bit. ‘Cause we no longer had to smell our way around the forest, you know? Those skills aren’t necessary. It’s scary to think about how if we offload some of our cognition to ChatGPT or to AI, that we might lose that capacity in ourselves. That’s something that I think is something I notice, is — I use a lot of AI myself — is instead of being good at coming up with a list of five things right away, I now just reflexively say, “Gimme a list of five things,” and my role becomes more of the executive, right? Making a selection between those, that’s a big change anytime you change human cognition. And so that’s one thing that think is, is certainly I approach with trepidation.
Maranda:
Yes.
Audience member:
You’ve talked a lot about ChatGPT, but what other platforms do you see are leading the way in just the quality of feedback that they’re giving?
Ben Talsma:
That’s a great question. So we see ChatGPT used most frequently. Like if you, I, I’ve got a graphic that I didn’t share here today that really shows it as the elephant in the room in terms of what we’re seeing in schools. It’s getting 80% of the use. We’ve got some schools though that are getting contracts to get a paid version of different sites, different sources. Copilot gets a lot of, of use right now. People are making use of Gemini as well.
Still, ChatGPT, I almost use it in the generic sense like Kleenex, right, or like, to Google. Just because when we interact with teachers, 80% or so of them are using that most readily. That’s a great question. So there are others out there. A lot of them, I think, are very similar in terms of how you put them to use. So, a lot of these strategies or ideas I think are, are tool-agnostic, right? They work just as well with any of the large language models or at least similarly well. And, so, it largely becomes a matter of preference then as far as what specific tool you want to put to use. Hopefully that answered the question. Wonderful.
Audience member:
Hi. I was wondering, you know, how many schools in Michigan would you say are actively taking advantage of AI resources? Like, you know, what you’ve provided?
Ben Talsma:
Well, that’s a wonderful question. So December of 2022 or so, right when this really started to enter the popular consciousness, we really started to invest in figuring out what this is, how it’s used. We had some webinars that were hugely successful, lots of people interested. We started to get some professional development with schools to help them build this. And when I asked people back then, not this school year, but the last school year “How much have you used AI tools?” One being never, five being implanted directly into my brain. It was mainly one’s and two’s all through last year, all through last year. One’s and two’s, maybe a couple of three’s. Every once in a while, you’d have an early adapter who was up there in the four-five range.
This year when we do AI work, Jamie is, am I right on this? You’re seeing lots more usage, lots more four’s and five’s. This feels like the year where every school is using it. Every school has at least some teachers who are making regular use of AI technology. Most schools are developing their policies and practices this year. So right now they, if you have a kid who’s in school, someone in that school is making use of AI technology for sure. And it’s becoming — this is the year I think where it’s really taking off in terms of prevalence of use amongst teachers. Is that helpful?
Audience member:
Yes.
Maranda:
That makes me happy because any student you talk to is using AI. So I’m glad that teachers are on board.
Ben Talsma:
Yes, yes, for sure.
Audience member:
Um question about, what is the thought about the appropriate age to start introducing this conversation into students’ vocabulary? ‘Cause they’re probably using it before we’re thinking about them using it.
Ben Talsma:
Wonderful, wonderful question. That’s still something I think that’s, that’s largely evolving with these tools. We have seen many of them are extremely text-based, right? And so, it doesn’t make a ton of sense for most kindergarten, first graders who are still developing that literacy to interact with these on a regular basis. Somewhere in that second to third grade range is when these tools might become accessible and when you wanna, I think, start having some grade-appropriate conversations. Things like, maybe exposing them to some of these tools that scaffold and supports that — just to give them a very structured, very limited interaction with that.
As they approach middle school, I think the temptation for things like plagiarism becomes exponentially more of a risk. And so, having those ethical conversations is something I think can start early, but really increases in importance as they enter that 10, 11, 12-year-old range. I’m making that up. I don’t have any kind of studies to back that up, but this is kind of what makes sense, I think, right now, out there for educators. So it’s a great question as well.
Maranda:
Let’s, let’s go here then here.
Ben Talsma:
Yeah, go ahead.
Audience member:
Are you seeing any kind of shift away from traditional exams to more project-based learning now because of AI?
Ben Talsma:
Now our, our sample size is a little bit biased there. The question is, “Are we seeing a shift more toward project-based learning?” One of the things that we really try and work on with schools is authentic instruction and project-based learning. So, the schools that we work with are showing a lot of that tendency. I’m not sure if that’s just a sample size bias kind of thing. But it is certainly something that’s being discussed more in the, the intellectual ether of education is, “Oh, could some of this PBL — could some of this authentic instruction be a way to counteract some of the pernicious effects of artificial intelligence in education?” So, I do think that’s something that is increasingly interesting to schools. Great question.
Audience member:
Okay. I’m trying to find the best way to ask this question. So, I love that this is such a great resource for our current educators. However, in the current stage, we’re already struggling to get teachers to become teachers and be part of the classroom setting. How is this going to be beneficial to our future educators? Because with all the resources, I’m not an educator, but I can easily have ChatGPT just create a lesson plan for me, and without any degree or anything, I can just teach a whole classroom. Obviously, we know as, all the certificates and all of that, you still have to go through your schooling. However, it just does not seem so enticing anymore when you are the one teaching children and helping them get through life.
Ben Talsma:
Interesting question. Interesting question. I think about that in a couple different dimensions. First of all, one of the main things that teachers complain about when teachers are complaining about things — they don’t get paid enough. That’s something that exists. The second thing is, they don’t have enough time, right? And so, to the extent that these tools help teachers be able to more effectively manage the load of work that they have, I think it has the potential to make education a more enticing profession, because you are less overwhelmed and more capable of doing the job that you’d like to do. So in that sense, I think they could be really helpful. In another sense, I think that a lot of people go into education because they love kids forming understandings, and they love kids, right? They love forming those connections.
And so by making us more efficient, perhaps with the information delivery side of things, where, theoretically, you could be getting some personalized tutoring from an AI system that’s helping you understand the specific things that you don’t understand, you can see more of those light bulb moments. You can facilitate more of those. But also, you can shift more of your attention to building those connections, to making those memories, to doing the kind of interesting hands-on learning. So I think it has the potential to make education more fun for teachers by allowing them to focus on those kind of rich hands-on learning experiences and on those kind of human connection elements of the profession in a way that still respects your time as a, as a human being. So I, I hope that was, did that make sense? Yes. Wonderful.
Maranda:
I’d say bring Ben in. ‘Cause I, I’m not a teacher, but I sat up there and I was like, “I’m excited about this. I can teach!” <Laughter>
Ben Talsma:
It’s one, it’s one of the best things. It’s when you’re in a professional element, you got a teacher in the back who crossed their arms like this, and it’s kind of like, they don’t like smartphones and they don’t like typewriters and they think everything should be in the tablets, right? They still have their chisel in the back. And then at the end they say, “Oh, I gotta, gotta tell you, I will try. I’m interested. I’m gonna try this out.”
This has potential because they all want to have an impact on kids, and if they can see some ways in which this can help them to do that, they get really excited. So this is, I think, my favorite PD that I learned — that I teach — because it has so many of those light bulb moments, those mind-blowing moments where teachers really see new possibilities for themselves.
Maranda:
Any other questions? What else? Do have one right here?
Audience member:
Full disclosure, I’m Ben’s dad. This is hopefully not a planted question, <laughter> because I didn’t like the one you asked me earlier. <Laughter> There are unfortunately school systems all across the country that are, I would say underperforming, right? The kids are not being educated the way we would like them to. Do you think AI offers a way to, you know, provide a real shift, a real step forward for school systems that are suffering?
Ben Talsma:
Yes. <Laughter> We like to think scientifically, right? We wanna see the evidence, we wanna see the data. This is the year I think that lots of schools are starting to figure out how to adapt this in the next two or three years. I think that most schools will be in a position where they’ve adopted a lot of these ideas about saving teachers’ time, allowing them to, to do what they do best, which is connect and teach their kids about — maybe helping these tools to help kids understand things they don’t understand.
So I think in the next two or three years, again, we’re getting to the speculative territory, you will start to see scores improve because of AI. You’ll start to see kids who understand things better because their teachers have AI, I would say within, you know, three or four years, you’ll start to see that creep of schools wrap their head around how to use this and then leverage it effectively to make an impact on students. So, you can come back in four years and five years and test that prediction out, but that is my sense, as teachers really integrate this into their practice, it does in a way — that I don’t think a technology in my educational profession or, or tenure has, has shown me — does have the impact the possibility of making an impact. Good question.
Maranda:
One from online.
Moderator:
I was a public-school teacher of a decade, but I taught general music. I’m wondering if you’re seeing AI used equally outside of what we think of as the core educational, like lit, math, reading, writing. Are you seeing AI being used in, by phys ed teachers, by music teachers, by art teachers, by foreign language teachers, et cetera?
Ben Talsma:
That is a wonderful question as well. Yes, we are seeing this used in really broad ways because it’s such a generalizable technology, right? One of the things that really got people excited about AI was we’ve always had AI. Microsoft Excel is an artificial intelligence program. Netflix’s prediction algorithm is an artificial intelligence program. They’re very narrow. They only do one set of things.
When we think about the AI tools, the generative AI tools that we have right now, they’ve got a much more general range of applications. And so, you’re an art teacher, you’re a music teacher, you’re a phys ed teacher, and you’re struggling to come up with an engaging way to teach a particular concept. You can get 20 ideas from AI and then you can converse about how to bring those to life, right? If you’re looking for different ideas, you’ve got ideas, you wanna get some feedback on them, you can get some feedback and then you can take that into consideration. You’ve got that wise mentor sitting right next to you. And so, you have the opportunity to really benefit from the insights of these technologies. We’re seeing them used in really broad ways across not just education, but other fields of disciplines, but certainly within education by all sorts of teachers.
Audience member:
Hi, I’m at the very, very bottom of knowledge of all of this. How do I start? Where do I go?
Ben Talsma:
That is a wonderful question. If you go, we’ll use ChatGPT as the kind of the model here. But they all, all the large language models kind of operate similarly. If you Google search them, you will find ’em. So look up ChatGPT, then you’re gonna log in using one of your standard internet authenticators, your Gmail address or your Apple ID, and there’s gonna be a box at the bottom, similar to what you saw when we were doing some of our ChatGPT modeling there.
All you have to do is type your request in that box and see what comes out. So again, one of the reasons that this has really exploded is because the learning curve is so shallow, they’re super easy to interface with, super easy to access, experiment and explore. And, so, the number one recommendation we have is just look it up, test it out. It’s a box where you type, and you interact with it and you see what it comes up with. So it should be something that if you, if you do that, when you go home today, you make some interesting discoveries about what it can do.
Maranda:
So, I would challenge you, get the app on your phone and then for fun, ask it about yourself. Give it your first and your last name and where you live, and you’ll be shocked at what ChatGPT can find out about you in 15 seconds. Then, ask it to create a ten-second video of your hometown, and you’ll be amazed. I mean, the information you can pull is exciting. So I love that. It’s like, just get started.
Ben Talsma:
Yeah, I mean, I, we love to teach kids to experiment, to explore and poke around with the limits of things. And that’s a great way to interact with these AI tools is try and stuff and try and come up with something that it can’t do, right? Try and look, look right at how it does a sonnet, right? Look at how it does a haiku about something that you’re passionate about. Challenge it to create some interesting pictures that you don’t think it could create pictures of. See what it comes up with. Make it make weird analogies. In what ways does the seven of diamonds face card — you know card from the card deck — teach us or reflect upon the modern political climate? <Laughter> It’ll make these interesting connections. Test it out, see what it can do. Great questions.
Maranda:
We have time for one more. Someone in the audience has a final question they’d love to have answered. If you think of a question, you can always email that to us. We’d be happy to answer that for you as well. You have one right here. Let’s do it.
Audience member:
This isn’t very important, but when I’ve seen this demonstrated before, I noticed that the person putting in the prompt always puts “Please”, and, like all these, like you’re interacting with a person. And my reaction is like, it’s not a person. <Laughter>
Ben Talsma:
Yeah, no, that’s a real thing. I think we actually finished with the most important one of all, is when you interact with these machines, I always say please, I always say thank you in advance, because if they do take over, I wanna be on the right side.
Audience:
<Laughter>
Maranda:
Thank you.
Ben Talsma:
Thank you all. Thank you.
Maranda:
If you want to learn more about what Ben and the entire Education team is doing, I would so encourage you to stick around. They’re going to be actually leading the tour of the Education Institute, which is right down the hill from here. So you will have to do a little walking, but it will be so worth it. You will meet an amazing tortoise and the incredible team. So you can head down there.
You can also stay and tour the first floor here and see what’s actually happening in the building. Our friends who are handling the microphones will be available to lead you on that tour. If you’ve enjoyed today and you’re saying “I want more of what they’re doing,” we want you to join us. Next week, on the 19th, we have a very exciting Public Lecture Series that will be handled virtually exclusive. It is on inflammation and depression.
The numbers for the people attending are through the roof because, as you know, we are, this is a, mental health is a huge situation and I love that we’re doing it virtually so that we can have some very private conversations on the side as our presenter, Dr. Brundin, is presenting. I think you’ll find it fascinating. So, there is still time to sign up for that. That’s on the 19th. We also have another one coming up on December 12th. You can find out about that as well. Check out the Facebook page, check out the website so you can be in the know with the Van Andel Institute. Thank you so much for joining us. Let’s give another big round of applause.
Image credit: VAI Learning Specialist Ben Talsma during his public lecture on how K–12 educators are using AI in the classroom.