The AI Tinkery: A sandbox for educators

Karin Forssell (00:00):

It's trying to get at that sort of sense of creativity and competence and communication and collaboration, all those things that we do in makerspaces, and bringing it into this realm of AI tools.

Dan Schwartz (00:16):

Today, we're exploring how creativity, tinkering and hands-on learning supports education. And can AI expand the possibilities for cultivating creativity in the classroom for learning and teaching?

Denise Pope (00:28):

It's super exciting, Dan. I know makerspaces are all the rage. Places that really give students the freedom to explore and to imagine and to design and experiment and bring their ideas to life. I'm really excited about this episode.

Dan Schwartz (00:43):

Well, we're lucky to have a guest with us who's at the forefront of creating the kinds of spaces that support this new kind of learning.

Denise Pope (00:52):

Yay. This is going to be a fun one. Let's start tinkering.

(00:59):

Welcome to School's In, your go-to podcast for cutting-edge insights and learning. Each episode, we dive into the latest trends, innovations, and challenges facing learners.

(01:09):

I'm Denise Pope, Senior Lecturer at Stanford Graduate School of Education and co-founder of Challenge Success. And I'm with my co-host, Dan Schwartz, Dean of Stanford GSE, and Faculty Director of the Stanford Accelerator for Learning.

Dan Schwartz (01:26):

I want to introduce Karin Forssell. She is the Director of the Learning Design and Technology Master's Program, which is a very, very successful master's program. Congratulations, Karin. And a Senior Lecturer here at the Graduate School of Education.

(01:40):

In addition to that, she directs the GSE Makery, which is a makerspace where students and faculty can explore how making things helps people learn. And maybe helps them learn something even beyond what they're trying to make. And now, she's been charged with making an artificial intelligence makerspace where you use generative AI and people come in and make poems for retirees, among other things. Welcome, Karin, it's good to have you here.

Karin Forssell (02:08):

Thank you. Good to be here.

Dan Schwartz (02:10):

Let me ask you, you were charged with making the AI makerspace, let's say of the AI. You made a very successful makerspace, it's on the first floor of our main building right now. There's always people in there. It's always noisy as hell.

Denise Pope (02:25):

That's a good sign.

Dan Schwartz (02:26):

Good or bad, I'm not sure. What did you do to make it successful?

Karin Forssell (02:30):

Well, it wasn't me alone, let's be clear about that. But we explored a variety of different things to try and get people in the door. We actually borrowed from museum education, the idea of threshold fear, and thought about what it would take for people to feel comfortable just coming in the door.

(02:49):

First thing we did was put in a doorstop, that didn't exist before, so the door is open. And then makery autocorrects to bakery, so we had cookies. And then we started to think about, "What are the other signals and ways that people can come in?" It's everything from that we have a video online to show people around the space if they want to check it out before they come. We have workshops, we have classes, we have open office hours, we have a manager who's there to answer your questions. A lot of students actually, we hire several students to provide extended hours and just to be liaisons to different programs and invite their friends to come. And then we basically say yes to whatever they want to make.

Dan Schwartz (03:27):

Denise is a novice. She hears the commotion and she walks into the makerspace for the first time. What happens to her? Do they say, "Here's a saw, go cut stuff"?

Karin Forssell (03:41):

The saw is in the back. Actually, what's up front and what a lot of people actually come in for is some of them come in because they want to fix something. Glue and sewing are actually top of the list. And then they get some ideas, there's lots of examples of different projects. And it's like, "Oh, I want to make one of those." And then there's some people who come in because they heard about us and they're in programs where it's like, "Oh, I heard you had another 3D printer. Can we use yours, because there's such a line at the other space?" We work a lot with the other makerspaces and we get a lot of referrals.

(04:12):

The first thing that happens though is you see the sign that says, "Welcome, please sign in in the book." And it's a bunch of different stamps, you get to stamp and put your name. And then somebody's going to come and say, "Hi," and, "How are you, and who are you, and what would you like to do here?" It's very personable.

Denise Pope (04:28):

Can I ask what a listener might be thinking right now? Why does a Graduate School of Education at Stanford University have a place for people to use glue and yarn and make stuff?

Karin Forssell (04:39):

That's a great question. There are lots of good reasons for this, and one is because they're big in K12 education. More and more schools are using a classroom to have 3D printers and laser cutters and vinyl cutters and a lot of robotic stuff. And have it all in one space with various curricula so that kids can explore the creativity that comes with making something by yourself.

(05:03):

Now, traditionally, that's happened in art, it's happened in industrial tech or shop classes. It's happening in cooking or family and consumer sciences, I think is their new name. It's happened in a lot of places. This is sort of a place where all of it comes together with a lot of computer facilitated making as well.

Dan Schwartz (05:20):

Denise?

Denise Pope (05:21):

Yeah?

Dan Schwartz (05:21):

Can you think of an activity where you make something and then it helps you do better on a school test or is making its own special track? You go to the makerspace to have fun making and maybe you learn a little persistence or you learn trial-and-error. Any way to squeeze science out of it?

Denise Pope (05:40):

Yeah, totally. Modeling, right? The first thing that comes to my mind would be modeling, or understanding geometry, which was very hard for me because it's flat on a page with numbers and stuff, but if I could see it, maybe. I also just remember, a long time ago, I made the planets. Probably everybody did that, right? With the styrofoam and the glitter or whatever. And to show how far away they are from one another. I think I could see all that. Yeah? Am I being too basic?

Dan Schwartz (06:12):

Karin? What do you think? What are the learning outcomes besides learning how to make the very specific thing you're making?

Denise Pope (06:17):

Or try something new and overcome and persist and problem solve, because those are good.

Karin Forssell (06:22):

Right, right, right. Well, my head was going in a slightly different direction, which is there are lots of benefits from teachers actually coming in and learning to make tools to help their kids learn because then they actually have to engage more with what kinds of maybe manipulatives, maybe other kinds of tools that they could help with a kid who has a real, maybe it's a common misconception or something that they're really struggling with and they could illustrate it or make some more tangible interaction.

(06:47):

You could do that peer-to-peer as well. You could have kids designing some kind of learning tool for younger kids or for peers or somebody in another class that would help them, by trying to teach another kid, they would actually be engaging more deeply in what is it that this is. And we've had all sorts of fun things built like a subject spinner for and English class. It's like, "Okay, you've got to do a story, so I'm going to make this little spinner tool and it's going to make you choose this actor and this mood and this situation," or whatever.

Dan Schwartz (07:20):

That's fun.

Karin Forssell (07:20):

Yeah, right? One thing is to get in deeper into the subject area. Another is to just sort of fuel the general idea of that learning is purposeful. In some sense, every subject is a making subject and that we are trying to have a particular set of tools to be able to create in the world, create arguments, create experiments, create solutions, whatever it is. And so getting into that sort of mindset of that we're here to solve problems.

Dan Schwartz (07:51):

That's interesting.

Karin Forssell (07:52):

I think would be transformable.

Denise Pope (07:53):

Okay. You just said something really that just went, "What?" to me, which is create arguments, which is right now in today's world, we've got a lot of problems with people speaking civilly to one another and to figure out ways to compromise and all that.

Karin Forssell (08:08):

Yeah, I don't mean arguments like create conflict. I mean laying out a set of facts in a way that leads to a particular conclusion. Thinking thoughtfully about what supports do you need to persuade somebody in a particular direction.

Denise Pope (08:26):

Give a super quick example of a makerspace thing that did that.

Karin Forssell (08:30):

You could create something on the laser cutter where you have all these different pieces and you had to slot them together as though they were a puzzle. And you look at how does this flow and does this actually lead you to a particular conclusion, or do you need to rearrange it in order to make this? Making an essay or something in the sentence level or the paragraph level, you could actually copy and paste. You can move things around physically and discuss what that means collaboratively, much easier than you can when it's on a Google Doc or something like that.

Denise Pope (09:00):

Super cool. Super cool.

Dan Schwartz (09:02):

I asked about the outcomes. I just want to say, not everything has to be justified in terms of outcomes. It's okay to listen to music and get to enjoy music just because it's music. It doesn't have to be justified for increasing the STEM workforce.

Denise Pope (09:16):

Okay, this is a crazy moment in School's In right now. I'm just telling you this because usually you are all about, "Why are we doing this?" And I can't imagine that parents out there are going to be super happy if their kids went to school just to listen to music or make some fun things in the makery. I feel like we've switched roles here, Dan.

Dan Schwartz (09:39):

I don't know if I want eight hours a day of kids making chess pieces with a lathe and glue and colored pieces of wood. I agree with that, but I just want to say some things are sort of their own justification. And so you don't always need outcomes. I think there is a place for beauty in the world, Denise, even if it doesn't make anything better.

Denise Pope (09:59):

My gosh, who are you and what have you done?

Karin Forssell (10:01):

All the art teachers are going to be happy to hear that.

Denise Pope (10:03):

Who are you and what have you done with Dan Schwartz from the previous episodes? Oh, my gosh.

(10:11):

Dan Schwartz, do you have a hobby?

Dan Schwartz (10:15):

Most of my hobbies turn into chores.

Denise Pope (10:19):

Like what?

Dan Schwartz (10:21):

I have a peach tree. I plant it. I fertilize it, and then pretty soon, I have to put a bird's net on it. And then I get all these peaches and I have to freeze it, and I sort of don't want to do it anymore.

Denise Pope (10:35):

Oh.

Dan Schwartz (10:35):

I have a short attention span.

Denise Pope (10:37):

I would call you a reluctant peach tree hobbyist.

Dan Schwartz (10:42):

That's very specific, Denise. Thank you. How about you?

Denise Pope (10:47):

I was thinking about this because I knew that we were going to be talking about hobbies. I feel like I'm sort of a boring person. I love to hike and be out in nature. That doesn't seem like a hobby to me. And I love to read also. I don't know. Is that a hobby? I have no idea. I like to cook, but I don't knit or collect stamps. When I think of a hobby, I think of things, and that's not in my world.

Dan Schwartz (11:17):

Okay. So you and I may need some empirical research to find out why hobbyists do what they do.

Denise Pope (11:22):

Yeah, yeah. But you've done some studying of hobbyists, yeah?

Dan Schwartz (11:26):

I have. And I don't promote my book very often, ever, but I'm going to read something from it.

Denise Pope (11:34):

Oh, boy.

Dan Schwartz (11:35):

We did a study where we asked all different types of hobbyists, like gardeners, people who brew beer, motorcycle racers, all sorts of hobbyists. We gave them a list of 25 possible things that would motivate them. To be liked, you might do a hobby to be part of a group, to feel a sense of control. Classic motivations. The number one is to see the fruits of your labor.

Denise Pope (12:04):

Okay.

Dan Schwartz (12:04):

Hobbyists like to create stuff. They make stuff, like "I brew beer and I've made this and I want to see how it turns out because I made it."

Denise Pope (12:15):

Okay, but how does that work with the motorcycle racer?

Dan Schwartz (12:18):

Oh, they're so tricked out. Have you seen how much gear they have?

Denise Pope (12:21):

Oh, so part of that is just showing off your thing?

Dan Schwartz (12:25):

Showing off is pejorative. Sharing.

Denise Pope (12:28):

With you, it would be to have someone taste the peaches from your tree? Yes?

Dan Schwartz (12:37):

Yes. I want to see how those peaches taste.

Denise Pope (12:41):

Right.

Dan Schwartz (12:42):

Right? And that's very literal to the fruits of your labor, so I like that.

(12:52):

Karin, you and I are going to have to do a little bit of work helping people's imaginations of what you can do, in case they haven't had opportunity to go to a makerspace to discover this. The Graduate School of Education is going to make an AI makerspace, but we're not allowed to use the word makerspace because everybody else on campus who has a makerspace, like engineering has lots of makerspaces, they disapprove because they think making has to be tangible. I look at them like Blockbuster, where they used to think watching videos had to be done with brick and mortar cassettes.

Denise Pope (13:25):

Oh, you're calling people old.

Dan Schwartz (13:27):

No, I'm just saying they've got to move with the times.

Denise Pope (13:30):

Okay.

Dan Schwartz (13:31):

But that's okay.

Karin Forssell (13:32):

This is more digital making, yes.

Dan Schwartz (13:34):

Yeah. My whole life's digital. Why does it have to be tangible?

(13:38):

Anyway, Karin, tell us what is it we're doing?

Karin Forssell (13:41):

Yes.

Dan Schwartz (13:41):

What are we doing now?

Karin Forssell (13:43):

We are starting the AI Tinkery, a sandbox for educators.

Denise Pope (13:47):

Very cute.

Karin Forssell (13:48):

Right. One of the differences between this space and others, just like our physical makerspace, is that it is focused on education, K12 and beyond, probably more beyond because we're including higher education as well. And it's trying to get at that sort of sense of creativity and competence and communication and collaboration, all those things that we do in makerspaces, and bringing it into this realm of AI tools, specifically generative AI tools.

(14:18):

You can come into this space and you can go ahead and play with stuff, try something out, come and find out what all the fuss is about if you haven't had any experience before. Maybe you come in for the candy and you stay for watching a video about a chatbot, and then you try out making your own. Or maybe you are going to go a little bit further and you have this question about, "How can I use this for this particular problem I have in the classroom?" Or, "I'm really scared about the ethics. How can I think about this and can you give me some examples of AI policies and what do we know from research?" A place to sort of aggregate all of that, and let people just explore, tinker with it.

Denise Pope (14:59):

I think it sounds awesome.

Dan Schwartz (15:00):

We had a session at a very large event where some of our cracker jack team presented how to make AI chatbots to teachers. And it was incredibly well attended. People would start by sort of saying, "Nah, this isn't for me." And at the end, they come running up to the teachers showing what they've made. It is the classic making thing where you really want to share and you're proud. I can imagine coming in and saying, "I want to create a lesson to help students compose pictures." How's the gen AI? What does the generative AI do for me there?

Karin Forssell (15:40):

Well, for one thing, you can quickly generate lots of examples. You could set them up and you could say, "What are you thinking about these different compositions? Which ones do you feel are more effective? Which ones are not?" One of the things that gen AI does for us is loads of examples really fast.

Dan Schwartz (15:58):

I would say, "Give me an example of a Malamute Labrador mix." And then I'd say, "Give me a different example." Something like that?

Karin Forssell (16:08):

Yeah.

Dan Schwartz (16:08):

Or I'd say, "Give me an example of a little girl who's lost in a friendly forest and meets a grizzly bear"?

Karin Forssell (16:14):

Yeah, absolutely.

Dan Schwartz (16:16):

"That has cheese and is playing a flute at the same time." And it can do that?

Karin Forssell (16:21):

"And it's purple." Yes. "And it's set on the moon." And it doesn't have to be realistic. It can generate these pictures in all sorts of different ways.

(16:32):

And by the way, you can also say, "In the style of so-and-so," Salvador Dali or something. You can get really specific. And one of my favorite examples of somebody who has done this is one of our colleagues with a three-year-old daughter. And they wrote a story together and the child can tell a story. And by the way, through natural language processing, she can just talk it and the computer will then transcribe it and then generate pictures to go with it. And then you print it out on a color printer and you've got a storybook authored by a three-year-old. That's really cool.

Denise Pope (17:06):

Okay, you got me there because I know not everything has to have outcomes, but the outcome there is potentially getting a kid very excited about literacy and about reading and about creating and understanding the connection between pictures and words. I came in dubious. I am leaving with a greater appreciation now of that.

(17:29):

And also something you said, Karin, that really struck me is this idea of learning how to collaborate and create and think through problems. Those are incredibly important skills. It's not just we're going in and we're making some sticker books. And it's not just actually, Dan, we're going in and listening to music. We're actually doing things to teach skills that kids need in the future, for sure.

Dan Schwartz (17:53):

Can I defend myself, Denise?

Denise Pope (17:54):

Defend. Defend away.

Dan Schwartz (17:57):

The great thing about the gen AI is that people really like to make stuff. Now, the challenge is can we have it so they get to make stuff that they really like to make, and it forwards their learning as well on things that don't necessarily always involve using gen AI?

(18:14):

A simple example is I learned to write by co-authoring with someone who was a better writer than I was. I would write at night and then he would take it and just completely rewrite it in the morning. Because he learned that editing is too slow. And then I would do again at night and I'd see all the moves he made. I can kind of do this with a computer now. And the hypothesis is that I'm going to be a better writer when I see how good it is at saying, "There are three points that we're covering." You can imagine someone who English isn't their first language and they get a chance to work with these tools and play with language. The trick is how can you make this sort of creative generation cash out in domain knowledge and skills?

Denise Pope (18:57):

I think there's another trick, which is how do you make sure you're teaching them skills for good and not skills for evil? Because I'm seeing a lot of the potential for deepfakes right? And for people who don't know what that is, creating pictures of a girl in your class without her top on. And so are you building in any safeguards at the AI Tinkery?

Karin Forssell (19:19):

Guardrails?

Denise Pope (19:20):

Yeah.

Karin Forssell (19:21):

I don't know exactly where we will wind up with this because we're sort of evolving as it becomes real. But I do think that educators are always going to be very interested in discussing the ethics. Obviously, there are lots of interest in the bias that comes from the training set being biased to start with. What kinds of issues might there be if students put in a prompt and get something that's totally inappropriate, disruptive, but also misleading? We definitely are going to see more tools, this is my prediction, over time. We're very early days right now. This is gold rush time. We're going to see more competent, more evolved tools over the next few years. And I think one of the things that we're going to see is more safe spaces for education that are not trained on everything on the internet, but on quality over quantity for what kinds of answers the AI can give.

(20:13):

We're going to see what kinds of limits it provides, but I think that in schools, we're likely going to have very targeted use of AI tools that help you to do particular things really well. Like compare and evaluate. I sort of suspect that our curriculum, it's sort of like when we got the calculators, remember when all the debates we had about letting people use calculators in school or on tests? Oh, my goodness. And now you're expected to because who does grocery store math in their head anymore? My mother did, yeah.

Denise Pope (20:47):

I think Dan Schwartz probably does.

Karin Forssell (20:48):

All right. With AI, I think we're going to find that we use these tools as helpers to get students to the place where they can recognize good work and they can go further, faster and level up with what kinds of writing they're producing. That's my hope.

Dan Schwartz (21:09):

I think we could have a very long discussion on policy. We don't want to make the same mistake we did with social media. But the theory of the AI Thinkery, colon...

Denise Pope (21:21):

Wait, it's a Tinkery?

Karin Forssell (21:21):

It's the Tinkery. Tinkery.

Denise Pope (21:21):

I kind of like AI Thinkery too though.

Karin Forssell (21:21):

That would be cool.

Denise Pope (21:21):

I know.

Karin Forssell (21:24):

Yeah. No, it's the Tinkery.

Dan Schwartz (21:32):

One of the hypotheses behind the AI Tinkery is that we need to get these teachers in there to see what these tools can do.

Denise Pope (21:39):

Yeah.

Dan Schwartz (21:40):

And to create a set of policies or things that students should learn, like personal policies or social policies. If you haven't experienced this and thought about how to use it, it's going to be too abstract. But I do think we don't want to make the same mistake we made with social media. We sort of want to-

Denise Pope (21:59):

Yeah.

Karin Forssell (21:59):

What would be really cool would be if all the teachers are the ones producing these new tools, because with AI, you can code like nobody's business. It's so much easier. And if we could get qualified teachers designing the new tools for schools, that would be amazing.

Dan Schwartz (22:18):

I like that, teachers aren't just using AI, but they're actually creating the tools that work best for their classrooms.

Denise Pope (22:25):

It makes so much sense, right? When teachers are the ones shaping the technology, they can make sure it's practical and impactful for learning. And it's a chance to avoid the pitfalls we saw with social media. If we bring teachers into the conversation early, we can set thoughtful guidelines and make AI a real benefit in schools.

Dan Schwartz (22:44):

I like the idea of both teachers and students using the same makerspace where they're both trying to learn how to use these tools. Maybe the students teach the teachers something about how to be a maker and vice versa. And so they get a chance to see what it can do, fire up their imagination, not worry about the outcomes, Denise. And there's a lot of potential here.

Denise Pope (23:06):

No, I love that. I love not worrying about the outcomes, you know that, Dan. I think it's really exciting stuff for the future of education, teachers, students, tinkering together. Very exciting.

(23:17):

Thank you so much, Karin, for such a great conversation. And thank you all for joining us on this episode of School's In. Remember to subscribe to our show on Spotify, Apple Podcasts, or wherever you tune in. I'm Denise Pope.

Dan Schwartz (23:30):

And I'm David Schwartz. Sorry, I'm just tinkering with my first name.

Denise Pope (23:34):

Oh, my God.

Dan Schwartz (23:35):

I thought I'd try it out in a safe space.