Exploring the Potential and Ethical Implications of AI in Education ~ 1038

Tim Villegas
Are you ready for the AI invasion? No, seriously, artificial intelligence will affect educators sooner or later—you might already have been affected. Before we roll out our Season 11 premiere next week, I wanted to share a conversation on the potential and the ethical implications of AI in education.

My name is Tim Villegas with the Maryland Coalition for Inclusive Education, and you’re listening to Think Inclusive, our podcast that aims to build bridges between families, educators, and disability justice advocates to create a shared understanding of inclusive education and what it looks like in the real world.

Michael Boll is a technology coach and educator who specializes in AI and its applications in education. He has a background in technology and has been providing professional learning experiences for educators on AI. Michael and I discussed the misconceptions about AI in education and how it can be used to benefit students. He emphasizes the need for educators to understand AI and its capabilities in order to create a positive narrative around its use. Michael also explores the potential of AI in personalized learning and supporting students with disabilities.

I think you are going to find this conversation very interesting. Also, make sure to listen to the end of the conversation—I’m previewing a new segment for Season 11 called the Mystery Question. Don’t miss it. After a short break, my interview with Michael Boll.

Tim Villegas
Michael, welcome to the Think Inclusive podcast.

Michael Boll
Thanks for having me. It’s great to be here.

Tim Villegas
We aren’t doing video, but if you were to see us, Michael is wearing a USC hat, and I’m wearing a Dodgers hat—yet neither of us live in California.

Michael Boll
Yeah, that is a little bit awkward, isn’t it? It almost makes it worse to show in the video. But I have a great face for radio, I’ve been told. So we’ll just leave it at that.

Tim Villegas
That’s fine. Michael, I invited you on because I want to talk about AI. But some people may be wondering: What’s your connection to AI?

Michael Boll
Well, I actually invented it—that’s now a big secret, and now it’s out on the podcast world. No, seriously, my background is mostly as a nerd, and that evolved into a technology coach. As a technology coach, you get to check out all the new things that come along and actually get paid for it.

I do professional development nowadays for teachers, so I’m always looking for new things to learn about. AI jumped right into the forefront because there’s so much stress, concern, and worry about it—especially after COVID—that suddenly teachers are going to be less relevant than they even were then. People think it’s the end of the world, that kids are just going to cheat, and on and on.

I guess all those things could be true if we’re not careful, but that’s what incentivized me to check it out. I want to share it with other educators so they can learn it, understand it, and then create and begin that narrative of how we’re all going to use it.

Tim Villegas
So you provide professional learning experiences for educators, and AI is something that you talk about. What are some misconceptions about AI that educators should know?

Michael Boll
That’s a good question. I think what’s so new now is that the concerns they have might become reality—maybe it’s going to take over the world, maybe it’s going to decide that we’re going to make paperclips and force us into indentured servitude. Who knows? That might happen.

The misconception is that students will only use it to cheat, that it’s going to revolutionize education in a negative way only. That’s why it’s important now to have these early conversations to help us understand how it works and how we can revolutionize AI to be something that benefits us and education.

You can imagine a virtual tutor that every student has, regardless of ability—helping them along, right by their side. If you’re in a classroom of 30, you have one teacher and then 30 instructional aides right there helping. But to make that happen, we’ll have to create structures and change our lessons to take advantage of that.

Right now, we’re in this disruptive flux where we don’t know how it’s going to go, and that’s where fear starts to take over—plus all the headlines saying it’s the end of the world.

Tim Villegas
I just read yesterday about how AI is going to be reading news for us.

Michael Boll
Yeah, I read that too—New York Times, NPR, and others are going to have it help generate news stories. I’ll take my writing, put it in there, and say “improve this,” and it’s a lot better when it comes out. I’m not a horrific writer, but I wouldn’t consider myself an excellent one. After AI, I’m pretty good.

Tim Villegas
So I guess we’re all going to become great writers. For years, I’ve subscribed to Grammarly—I’m sure you’ve heard of it. It immediately made me a better writer. It used to just do grammar, but now there are AI tools built in. You can write a paragraph, highlight it, and say “shorten it” or “improve it.” It’s become so much easier to write.

That’s my job—communications. Anything that helps me do that better and faster feels like a win.

Michael Boll
Right. AI in its current form is perfect for helping those of us who already learned how to write. But the concern is: What about those still learning? Are they just going to put it down and remove all the learning? That’s definitely possible today.

If you’re a teacher who gives a lot of homework and doesn’t create structures in class where students do it step by step, that’s going to happen. The incentives are huge. If I’m a high school student who wants to get into a good university, I’m going to use those tools.

Tim Villegas
Do you see a good parallel to AI as the calculator? I had a conversation recently where someone said the calculator was “dangerous” when it first became popular because kids wouldn’t learn math. Now we have calculators on our phones—or even on our watches.

Michael Boll
Right. Back in the day, calculator watches were a thing—nerd credentials here! But it was an isolated incident for a specific course, so it wasn’t super difficult for teachers to adapt. It’s been a healthy change because now we focus more on the structure and concepts of math, and then use tools to assist us.

With AI, it does everything on a wide scale with super easy access. We’re on laptops all the time, so it’s hard to tell whether students are using AI or not. If we follow the calculator model, it could be great—but there’s going to be a lot of disruption initially. Teachers have to understand how it works, and that’s harder than understanding a calculator.

Tim Villegas
What is the best way to understand what AI can and can’t do at this point?

Michael Boll
The first thing to do is just get an account. You can use ChatGPT, Google’s version, or Microsoft’s version—those are the main three. They have different levels of quality. Start asking questions.

For me, a lot of it is brainstorming: “Come up with 10 names for this course,” or “Research Tim and come up with five questions I should ask him.” That makes me smarter and helps me understand ChatGPT’s ability.

Then I might say, “I need a lesson on dinosaurs for second graders. Make sure you incorporate basketball because we’re doing a basketball unit.” Or reverse it: “I’m a PE teacher and want a cool game that involves dinosaurs.” It’ll generate ideas, and if I don’t like them, I’ll say, “Come up with 10 more.”

As you use it more, you realize its power. Think of it as: What am I doing today? Can AI help me? Often the answer is yes, and the responses can be delightful.

Perhaps the biggest beneficiaries will be elementary school teachers, where students aren’t on laptops much. They’ll use AI to be super creative in everything they do.

Tim Villegas
What about things AI can’t do right now—but people think it can?

Michael Boll
Good question. AI is generative—it doesn’t just research the internet. It generates content based on massive algorithms that understand human language. It doesn’t really know what it’s saying. If it doesn’t have a reference point, it’ll make stuff up.

That’s a concern when you ask factual questions—you have to fact-check, which is a pain. For creative tasks, like lesson plans, it’s great.

On the factual side, it hits a wall. For generating images, it’s not that good yet. Music? Okay, but not great. These will improve over time.

Tim Villegas
There are these AI-generated videos—Star Wars or Lord of the Rings made in the style of a Wes Anderson film. Did you ever see those?

Michael Boll
No, but I can imagine.

Tim Villegas
They were fantastic. I love a good Wes Anderson film, and I also love Star Wars. When that first one came out, I thought, “How did they do this?” Then I realized it was AI. I couldn’t tell you how it works, but I know it does. We’re at the beginning of this whole thing. I can’t imagine what it’ll look like in one or two years—let alone 10.

Michael Boll
Exactly. I used to teach a film class. Soon, I’ll be able to tell iMovie, “Make this film look like a Wes Anderson film,” and it’ll do all the cuts, music, and style. As a director, I can decide whether I like it or not—or choose which artist I want it to emulate.

Think about editing: I’ll have options to instantly convert something and then use that as inspiration for my own creative work. You’ll even be able to star in the movie—AI will deepfake you into a character.

Before coming on today, I was listening to a podcast called Hard Fork from the New York Times. They talked about a Netflix show that uses deepfakes. It’s a reality show where people watch their spouse cheating—but it’s actually a deepfake. They don’t tell them in advance, so they capture all the reactions. Horrifying idea for a show, but it’s happening.

Tim Villegas
That’s like a Black Mirror episode.

Michael Boll
Exactly—probably one of the best ways to understand AI.

Tim Villegas
One misconception is that you can just ask AI to generate something, copy and paste, and you’re done. But since I’ve been using AI tools, I’ve never done that. It’s more like a first draft. You take what it gives you and refine it—like a director deciding what to keep.

Michael Boll
Right. And then there’s prompt engineering. When you get that first result, you can say, “Make it shorter,” “Make it punchier,” “Make it happier,” or “Make it sadder.”

For example, with the Common App essay: You can tell ChatGPT, “Answer the Common App question, reference Elmo, mention that I was captain of the volleyball team, worked with individuals with diverse needs, and had a job at an ice cream shop.” It’ll intertwine all that into a solid essay.

That’s just ChatGPT 3.5. If you pay for GPT-4, it’s even better. Then you modify it. Using prompts to get better answers—or refine them after—is key.

Tim Villegas
So is that cheating, Michael?

Michael Boll
I think it’s cheating if you don’t cite it.

Tim Villegas
Let’s say I use your example: I get a great first draft from AI, modify it a little, keep 75–80% of what it produced, and submit it without saying it was created by ChatGPT or Bard. How do we feel about that?

Michael Boll
It’s like asking you, Tim, to write me an essay. I give you the prompt and details, you write it, and I keep 70–80% of it. That would misrepresent my writing ability. It doesn’t reflect who I am as a student.

But the incentive to misrepresent yourself is massive—getting into a top university, earning scholarships. That’s why we’ll have to change structures. We might have to eliminate essays because readers will always wonder if they’re legitimate.

I hate reading stuff I know is from ChatGPT. Before, I was getting to know someone; now I’m just reading words on a page. So yes, there’s a serious ethical issue. And as long as the benefit outweighs the moral discomfort, students will do it.

Tim Villegas
So is this the major problem we’ll have with AI—how to cite it, whether it’s ethical to use it?

Michael Boll
It’s ethical if you cite it, but you can’t cite a friend or AI for your essay—it has to be your original work. So how do we create classroom situations where students are assisted along the way, rather than told, “Go write this”?

Unless assignments are broken down into pieces that are checked along the way—or we eliminate homework—you’ll see a lot of this. Lectures on ethics will help somewhat, but incentives drive behavior.

Tim Villegas
As far as how we can use AI to support learners with disabilities—have you put much thought into that?

Michael Boll
Yes. I was preparing to develop a course using a UDL (Universal Design for Learning) model to help students across the board. Currently, students can take their writing, put it into AI, and ask for feedback on voice, grammar, or sentence structure. That helps everyone.

Right now, you have to do it manually, but companies are starting to create tools that structure this better. For example, Khan Academy is launching Khanmigo, an AI tutor that assists students along the way. It’s designed to help—not cheat. It pushes students with questions, provides guidance, and ensures the instructor knows the student went through the process.

The potential here is massive. AI could help teachers understand where each student is academically. Imagine a classroom of 30 students with different needs—you can’t realistically know every detail. AI can identify gaps and personalize support.

Tim Villegas
I like how you frame that. We need to change what we expect from learners. If the assignment is just “Write a 500-word essay on the Civil War,” that’s a problem. But if we expect some built-in help—whether through AI or something else—that’s more helpful.

Michael Boll
Exactly. AI can figure out where students are and help teachers create targeted interventions. Personalized learning has always been the dream, but it’s hard to scale. AI could make it possible—assessing students, guiding them, and giving teachers dashboards to track progress.

Teachers could then create fun, community-building activities while knowing which students need extra support. That’s exciting.

Tim Villegas
Something came to mind when you were talking about personalized learning and an application knowing students well enough to offer suggestions to the teacher or the student. That’s already happening with the data we give to social media companies. I’m scrolling through Facebook or Instagram and I get ads for audio and video editing software or musical instruments.

Michael Boll
Or how to be a better husband—that’s my wife or husband or something.

Tim Villegas
I remember a Meta survey asking if I like personalized ads or not. I have a mixed reaction. On one hand, I like seeing ads for audio equipment instead of things I don’t want. On the other hand, I don’t like that it knows me.

Michael Boll
It’s saying, “We’re going to collect all this data on you and know you no matter what. Do you want personalized ads that show we know you, or do you want us to be more subtle?” Obviously yes, because they’re not going to turn the switch off.

Their goals are to make money, compete for your attention, and get you addicted to screen time so you constantly pick up your device. That’s the incentive behind everything they do. Outrage releases cortisol in your system and keeps you on those platforms.

Hopefully, as educators, that’s not our goal. Our goal is to help you become the best person and most effective learner you can be. So collecting that same data at a large, aggregate scale—and scaling it across a middle school with five or six different teachers—is awesome. It’s also a study: as I serve up useful things—

Tim Villegas
Do you see a time in the future where educators could, based on student profiles, schedule a whole year’s worth of lessons—a curriculum?

Michael Boll
Hopefully not. Even if I’ve been with a student for a month, I don’t really know what they’ll be like in six months. I’d have goals for where I want students to get to, based on their abilities and grade level, and then AI would help along the way.

Take music class. Maybe I’m terrible at eighth notes. AI can listen to my playing and produce exercises to help me improve. Because it’s AI, it can be creative—use songs that are fun for me—so I learn in a way I enjoy while picking up the skills I need. That’s better than a band teacher with 150 students doing the same music for everyone. Maybe that reaches 50–60% of the class, but there’s still 40% that needs more specific help. AI can scale, and a teacher can’t.

Tim Villegas
It’s going to change how schools are run. We’re recording this in the summer of 2023, and educators are at home or in pre-planning, scheduling out their year: “In September we’ll cover this, in October we’ll cover that.” We may not have everything planned to the minute, but there are basic structures. What I’m hearing you say is that maybe we won’t need to do that. Maybe it’ll be more responsive to student needs.

Michael Boll
I think it can be. The question is whether we’ll do it. If I’m a math student and we learn quadratic equations in April, and I still haven’t learned it, the current system says, “Tough luck—on we go to the next thing.” Or, “Work harder. Get more tutors,” and add to the stress you already have.

If we leverage AI, we can individualize and aim for mastery—actually learning something—rather than sticking to a rigid calendar. Khan Academy has learned that students might be slow in one area and much faster in another. They catch up. Sometimes you just need more time in one area and less in another.

We’ll need to divorce ourselves from the idea that we do a certain thing in April and another in May by topic. Sure, we might cover the Civil War in May, but AI can help carry writing conventions from the Revolutionary War unit into the Civil War unit.

Not everything should be individualized time. We also need to come together as groups. We don’t want kids just sitting at their screens with AI nagging them. We want them to work cohesively as a group and understand their role in society.

It’ll be a big disruption. We won’t get it right in the beginning. We’ll make mistakes. We should approach it with a “minimum viable product” mindset—good enough to start, with the assumption that it’s not actually good enough—and learn and improve as people use it. There will be backlash because people feel fear and say, “I just want to go back to basics,” which is something they understood. But you can’t just turn the internet off.

Blocking AI? Go ahead and block it. Students will get around it—maybe not at school, but they’ll use it at home when they have no guidance.

Tim Villegas
I subscribe to Education Week, and I have to show you this headline. It says: AI Invasion: Weighing the Benefits and Drawbacks of Artificial Intelligence in Schools.

Michael Boll
There you go—clickbait on actual paper! I could use ChatGPT right now and come up with at least 15 better title ideas. I can’t believe they didn’t.

Tim Villegas
I haven’t read the article yet, but I’m sure it’s about fears and benefits.

Michael Boll
Exactly. Probably half the people are excited, half are worried. And then a quick summary at the end about what we should do.

Tim Villegas
Maybe I wrote it—you never know.

Michael Boll
Maybe AI wrote it! That’s the fear—that AI will replace journalists. Same with teachers and lawyers. There will be teachers and lawyers who use AI and those who don’t—and the ones who don’t will all be unemployed.

Tim Villegas
We’ve talked a lot about the benefits and challenges of AI, and how it can help learners with disabilities and all learners. We have a lot of educators listening. What’s the big takeaway you want them to have from this conversation?

Michael Boll
Learn it. Start using it. Find a cohort of people who want to create ways to use AI to benefit education. Help create that narrative.

Know that in the beginning, we’ll be terrible at it. It’ll do things we don’t want it to do. Kids will use it in ways we don’t want them to. That’s normal. Be prepared for that and adjust as we go forward.

If we don’t have an open-minded philosophy, we’ll succumb to fear and try to block things. Then AI will move forward without our guidance. Let’s guide it rather than letting it go on its own.

Tim Villegas
Having an open mind is so important. The things we fear and try to stop just end up having a life of their own.

Michael Boll
Exactly. It’s like a slow flood coming. You can resist it for a while, but eventually, it’ll get you. Better to build the boat now.

Tim Villegas
This has been a really fascinating conversation, Michael. I have one more question for you if you’re up for it. This season of Think Inclusive, I’ve been doing something called a Mystery Question. Are you up for it?

Michael Boll
Of course. I can’t say no.

Tim Villegas
Okay, I have a stack of prompt cards. I’ll select one, and we’ll both answer it. Here we go: Would you accept a fatal mission in exchange for a lifetime of support for your family?

This is an awful question—the worst one I’ve had so far. Congratulations, Michael.

Michael Boll
No, I wouldn’t. I’m hopeful—maybe arrogant enough—to think I can help my family survive and do well without killing myself to do it.

Tim Villegas
I agree. I would not accept a fatal mission either. So there you go, all you kidnappers out there—don’t even try it.

Michael Boll
Next time, let me go through the cards first!

Tim Villegas
I know. I should’ve asked ChatGPT for a better question. Actually, let me try: Write a random question to ask my podcast guest. It says: What has been the most profound or transformative moment in your life, and how did it shape the person you are today?

That’s a good one.

Michael Boll
For me, it’s having a son with profound autism. That’s been the biggest change in our lives. It’s caused us to change our lifestyle and how we help both him and his sister. It’s made us more creative than we might have been otherwise.

Tim Villegas
For me, it was working with young children with autism right out of college. I had no idea what autism was, but that job pushed me into education. It was transformative.

Michael Boll
Thank you for making those choices and continuing to do this work. We need you.

Tim Villegas
Thank you. It was my pleasure to have Michael Boll on the Think Inclusive podcast.

Thanks for listening to this preseason bonus episode. I’ve dropped some links to things Michael and I talked about in the show notes. Can’t get enough of Think Inclusive? Become a patron at patreon.com/thinkinclusivepodcast.

Follow us on social media—and if you follow us on Threads, I’ll post some dad jokes written by AI. Let me know what you think.

Find us at thinkinclusive.us. For more information about inclusive education or how MCIE can partner with your school or district, go to mcie.org/contact.

We’ll be back next week with our Season 11 premiere. Enjoy the rest of your week, and remember: Inclusion always works.