AI in schools: Perspective from a Minnesota teacher turned AI education consultant

NINA MOINI: As students and teachers are settling into their school year routines, we're going to take a look at how artificial intelligence is changing teaching, learning, and well-being in the classroom. This week on Minnesota Now, you'll hear from a teacher, a student, and a mental health professional about how they're viewing AI's impact on young people.

To set the tone for us, we're going to turn to Jon Fila. He was a finalist for Minnesota's Teacher of the Year Award back in 2022, but now runs Inclusive AI Strategies, which provide school districts with guidance and educational tools around AI. Thank you so much for your time today, John.

JON FILA: Hi, Nina. Thanks for having me.

NINA MOINI: Jon, you were a teacher, I read, for 26 years. Would you tell me a little bit just about what you taught, where you taught, and how AI came to be on your radar?

JON FILA: Oh, yeah. Sure. So I started teaching at a traditional public school right out of college, and then quickly discovered that I was more interested in different types of settings in student populations. I taught in a private Catholic middle school for a year, and then moved to more alternative settings in the metro area--

NINA MOINI: Mm-hmm.

JON FILA: --and working with students with behavior issues, drug and alcohol recovery, pregnant and parenting teens, just about any type of student population, and then worked with districts to have unique programming that suits the needs of those types of students. And then, at some point, I gravitated towards online learning, and spent a lot of time in digital accessibility in that space.

And about three years ago, I started to notice some really strange student submissions starting to come in. And I felt like it was in my best interest and my student's best interests to try to figure out what was happening with these AI submissions that I was receiving. And then it just kind of went from there.

And I was seeing more, and more, and more of it happening. So I started writing about my experiences, which turned into a few books on the topic. And developing what I thought were some insightful and unique strategies for dealing with not only student use, but how we could use it as educators. And what does appropriate use look like? And what would be an example of inappropriate use?

NINA MOINI: Yeah. So you mentioned there, perhaps, some students-- and I've heard this from other educators-- submitting things that are not written by them, maybe a ChatGPT kind of situation. But I would love it if you could describe what AI in the classroom currently looks like.

Obviously, students have tools that they use on their own computers. And then you also mentioned it could be ways that educators can use the tools. But what does it look like in the classroom right now? Because I think AI really sounds very distant and like a bunch of robots, at this point, to a lot of people.

JON FILA: Well, I think it looks like a lot of different things, depending on the teacher, depending on the situation and the setting. So that could be people that want to ban it outright and not allow it at all to maybe the single biggest disruption that we've ever faced in the field, where we've got this thing sitting there that people are tempted to use because it provides a quick and easy solution to so many traditional types of tasks and solutions.

So we don't sit with questions very long, since search engines became available. And so, now, we have something that can provide those types of responses even faster and maybe more articulate than a search engine might. So it's a struggle because there isn't clear guidance in so many places on how we use it, and where it integrates with what we're already doing, and with what our objectives are.

I would say it was a bigger struggle for me to adapt to that in the class than it was the pandemic response with the emergency remote learning and all that, the challenges that came along with that.

NINA MOINI: Mm-hmm. You're just saying everything--

JON FILA: It's more frustration right now.

NINA MOINI: --yeah, everything that's at the fingertips for students that isn't just looking through a book like it used to be. You mentioned people who are just wanting to ban this outright in the classroom. What are you thinking is a reasonable way to approach AI in the classroom?

And do you think it's reasonable to do away with it? And when you're in the class, you're going to do things the way we always used to do it. Or what are you seeing?

JON FILA: Well, I think to ban it is a little bit shortsighted. And we create new inequities when new technologies like this come out. And so students-- there are going to be students that get around bands.

They have their own devices. They have their own data connections. They'll find a way. When people are overworked, educators and learners together, they're going to look for ways that they can take a shortcut.

So I think it's not realistic to suggest we just can't use it at all. These tools exist. They're easily accessible. Many of them are free. And so I think finding ways to integrate them appropriately make more sense than trying to bury our head in the sand.

NINA MOINI: Yeah. And what do you think or hear from students about using AI. Because I feel like students want to be at the table, too, in deciding how their education is structured.

JON FILA: Well, that's really interesting because, yeah, I've done quite a few student surveys over the last few years. And surprised to see that a lot of students don't see it as plagiarism. They don't see it as cheating.

NINA MOINI: Oh.

JON FILA: Asking AI to create a response for them, they think it's part of their natural workflow. And if they're doing the prompting, and they're making their adjustments, they see it as their own unique, original work, for some reason. So there's a real disconnect, I think, in how students that are growing up with this now see it versus how we might see it from a more traditional educational experience.

NINA MOINI: Is there an example that you might want to share of maybe somewhere that you have seen AI being integrated responsibly, maybe in lesson plans or in assignments. What would that look like to you at this point?

JON FILA: Yeah. So there are a lot of tools that are kind of sold as click it and get a lesson. And that's what I tend to be a little more skeptical of, because of the bias that's associated and implicit within these large language models, like ChatGPT, or Cloud, or Google Gemini.

NINA MOINI: Mm-hmm.

JON FILA: I think it's important that we understand that those biases are a natural part of that. I actually call large language models bias engines. But I do think there are ways to integrate it responsibly to make things like special accommodations, improving accessibility of digital content, aligning content to standards, making improvements to lessons, finding patterns in students' performance, using it as an assistant or a collaborator. There are a lot of ways we can do that.

And there are strategies for avoiding the bias that comes with these tools. But if we're looking to some of the solutions that are out there, that guidance isn't part of the tool. We have to know how to do that as the user, and be aware of how they work, how they're creating their responses, and then being able to analyze them with our own expertise to know if it's really doing what we're asking it to do.

NINA MOINI: So, obviously, realizing every situation is individual. But it sounds like what you're saying is a rule of thumb is if it's AI that helps with equity versus not helping with equity, and putting more bias into the mix.

And you also mentioned something, Jon, about how people don't really have a lot of regulations around this. It kind of reminds me of social media, which started to come on the scene. And then years later, we're still talking at the federal level, at all levels about how to regulate it and how to make it safe.

What do you think needs to happen with AI right now to try to get ahead of some of the issues that are already popping up? Is this a legislative thing? Is this passing laws? Is this individual schools? What do you think?

JON FILA: Well, I guess in our current climate, I'm terrified about regulation. Because, in my view, when I'm using these tools, I'm asking it to incorporate a diversity of perspectives. I'm asking it to incorporate gender, race, location. I'm asking it to be more inclusive, right?

And so if we're creating regulations around that don't include guidance on that type of use to avoid the implicit bias, then what are we really regulating? What are we asking it to do?

The tools are going to sound like us. And if we're biased, that's a problem, if we're not sufficiently analyzing it. So if we try to create regulations, who's creating those regulations? What type of expertise do the people doing the regulations have on those topics to actually make something that's suitable for student use?

We don't really want these tools teaching history, for instance, in ways that we know have already been debunked. If a teacher is creating a lesson, these tools will generally suggest things that address things like student learning styles, or left, right brain thinking, things that have been thoroughly debunked by research now for decades.

But that's what we get in the responses because the overwhelming training on these tools has reference to those kinds of things. So you really need a foundation in what does good teaching look like to be able to evaluate the types of responses you're getting.

NINA MOINI: Yeah. So there's regulating how the AI tools work. And then there's also regulation around how they could or should be used within the classroom. Fascinating stuff, Jon. We're going to explore it more this week, as folks are getting back to school. And we hope you'll join us again sometime. This is going to keep being--

JON FILA: I'd love to.

NINA MOINI: --a thing. So thank you so much.

JON FILA: Yeah. Thank you.

NINA MOINI: All right. That was Jon Fila, a former teacher, and now the owner of Inclusive AI Strategies.