AI in the English Classroom, with Amy Onstad

In this episode of Unpacking Education, we explore AI in the English classroom with veteran educator Amy Onstad. With more than three decades of experience teaching high school English, Amy shares how artificial intelligence is reshaping both teaching and learning, from helping teachers save time on tasks like writing recommendation letters to supporting students as they prepare presentations, brainstorm ideas, and refine their writing.

At the same time, Amy reminds us that AI is a double-edged sword. While it can enhance feedback and streamline workflows, it also raises important questions about academic integrity, trust, and how students will be evaluated in college and beyond. Through practical examples and honest reflections, Amy discusses how clear expectations, strong relationships, and open conversations with students can help educators navigate AI responsibly while preparing learners for a world where AI will be part of everyday life.

Visit AVID Open Access to learn more.

Transcript

Amy Onstad 0:00 Whether students like to use it or not, they're going to be in an AI world, and the jobs are going to have AI. Even though it's not a standard, we need them to be ready to tackle AI. To withhold it and make it a source of angst and anxiety — "you're in trouble, and I caught you" — I think would do them a terrible disservice.

Paul Beckermann 0:21 The topic for today's podcast is AI in the English classroom with Amy Onstad. Unpacking Education is brought to you by AVID. AVID believes in seeing the potential of every student. To learn more about AVID, visit their website at avid.org.

Rena Clark 0:38 Welcome to Unpacking Education, the podcast where we explore current issues and best practices in education. I'm Rena Clark.

Paul Beckermann 0:49 I'm Paul Beckermann.

Winston Benjamin 0:50 And I'm Winston Benjamin. We are educators,

Paul Beckermann 0:54 and we're here to share insights and actionable strategies.

Transition Music with Rena's Children 0:59 Education is our passport to the future.

Paul Beckermann 1:03 Our quote for today is from a new study by the Center for Universal Education at Brookings titled A New Direction for Students in an AI World: Prosper, Prepare, Protect. The authors pose this question: as educators contemplate integrating AI into teaching and learning, how can they embrace its potential while preserving human agency, deep learning, and pedagogical judgment?

All right, there's a big question for you two. Rena, Winston, what are your thoughts?

Winston Benjamin 1:37 I think it's interesting to see how people are afraid of technology taking away the art of teaching. Technology provides a lot of opportunity to do things more simply, but just as when we had computers integrated, just as when we had other technologies integrated in schools, it's valuable for teachers to realize that they are the users of that tool. This provides a strong basis for teachers to remember how powerful they are as they utilize a new tool — that they can make pedagogical judgments in order to provide deep learning for students.

Paul Beckermann 2:27 Thanks for those insights, Winston. Rena, what are you thinking?

Rena Clark 2:30 It's interesting. I'm actually thinking of this as both an educator and a parent. I am literally supporting my oldest son with geometry right now. I have not taken geometry in a while, and without AI, I would not be able to provide the support he needs. I think about that leveling of the playing field — how are we able to still have kids think deeply, but also provide supports that can really help them?

Here's how we're using it: I take a picture of some of the geometry and we ask it, "Can you explain what this means? How do we learn? What strategies do we apply?" Then we ask for a similar problem, then we ask it to look at our work. It is life-changing right now, because it's allowing us to support him in a better way and jogging my memory from — we won't say how many years ago — since I've had geometry.

Paul Beckermann 3:43 Good thing he didn't have geometry four years ago.

Rena Clark 3:47 You know what? We would have figured it out. But so many kids don't have a person to do that for them. How are we supporting kids that don't have that — kids who are just supposed to go home and figure it out on their own? It's a really powerful tool in that way. So how can we find that balance?

Paul Beckermann 4:07 I love that. We get to dig into the topic a little bit today, and to help us with that conversation, we have Amy Onstad with us. Amy is an English teacher at Zimmermann High School and an adjunct professor at the University of St. Mary's. Welcome, Amy.

Amy Onstad 4:23 Thank you. Hello. Good to be here.

Rena Clark 4:26 Amy, we need the English perspective — I've been talking about math.

Paul Beckermann 4:31 Now we get the insights into English. Do you want to take a minute and tell our listeners a little bit more about yourself?

Amy Onstad 4:41 I would say one thing: in the world of tech and all of the things, let's not forget about books. I am reading all of the time. My book count right now is 79 books this year. I know — do the math on that one, Rena.

I appreciate the technology that I have and that my district provides, and I appreciate that I have Pam. Everybody needs a Pam in their life — that's my big advice. I love to come home and not do any of that. I love all kinds of reading and talking to kids about reading. I've been in the same district my whole career, so this is year 33, and I'm getting excited about exploring retirement with my financial advisor. That's where I'm at right now.

Paul Beckermann 5:57 And you mentioned this before we started — what are you teaching right now?

Amy Onstad 6:03 I teach both AP English Language and AP English Literature, and I teach Honors 10 English and Senior Composition, which is a regular English 12 class.

Winston Benjamin 6:15 Thank you so much for helping ground our audience in who you are. Here's a big-picture question: what do you see as a benefit of AI in education as we continue to move into the tech world?

Amy Onstad 6:30 Would you like the benefits from the student angle or from the educator angle?

Winston Benjamin Both. Why not?

Amy Onstad Okay, why not. We all know it's a double-edged sword — there's the good and the bad. But what I've seen from students is they're very clever with how they're using it.

I write 25 to 30 letters of recommendation per year, and fall is usually a big time. I've started using AI as a jumping-off point — students fill out a form for me, I put it into AI, ask it to start crafting the letter, and then I go back and take away the things I don't like and personalize it. Well, I had a student who took my letter and put it into his own AI. He has set up his own chat to be personalized for his personality, which is very blunt and straightforward — "don't give me 'oh, it's so nice, you're so good,' just tell me." He put my letter in there and asked the AI how I could improve it, then sent me back the chat showing how I could fix my letter for him. I'd just spent an hour on the letter — but that's a student using AI to go after what he needs.

I had another student do a presentation who put the rubric into the AI and said, "Listen to my presentation and tell me how I'm doing." He performed it and got feedback before he got to me. Those are students who are learning how to use it really well.

As for literature — it is not accurate, and it does not understand the nuance yet. Students will ask it about plot things and it's just wrong. Poetry and literature, it does not do well. But for writing, it's very helpful.

Rena Clark 8:49 And on the educator side — you talked a little about how you used it for letters of recommendation. Are there other benefits you see?

Amy Onstad 9:01 Yes. I've started using it as a jumping-off point for multiple-choice quizzes, because AP students do have a multiple-choice section on their test. I hate writing those, but I know students need them, so I use AI to start and then refine from there. It's still time-consuming, but far less so than creating everything from scratch.

I have teacher friends who use it to create rubrics. I've had mine in place for so long that I don't need that, but if I were a new teacher coming in, I would definitely use it for that. Others have used it for anticipatory sets and discussion questions, which are usually pretty decent.

I have also used it to grade argumentative papers — I'll put in their paper and the AP prompt, do my own grading, and then tell the kids what AI said. Then we have conversations about whether we trust AI on that or not. Those are some of the things that have been helpful.

Rena Clark 9:59 I'm curious — I worked with a lot of AP high school ELA teachers last year supporting them with AI, and I heard a lot of different perspectives. What are the potential challenges or drawbacks that you've faced, or that you've seen others face?

Amy Onstad 10:21 It will come as no surprise that the big one is using AI to write the paper. The other challenge is that the tool we use, Turnitin, is not reliable in determining AI usage — it's just not. GPT Zero is a little more reliable from what I've seen, but that has created interesting conversations with students. When I get a paper showing up at 45%, 55%, or 100%, I put a note on it and say, "Check AI — this is showing this." I get emails instantaneously. Students come and find me, their chins are quivering, they're nervous: "I swear I didn't." In that way it's led to great conversations.

My biggest concern is for students who have been writing all their own papers, and their high school English teacher says it's coming up 40% AI. When they go to college — and I don't know if you've heard stories online — there are students saying, "I did all my own work," and the professor says, "It's AI, F." There seem to be some very stringent things in higher ed, and I'm worried for my students going forward. There are tools like Brisk that help, but I don't know what to tell them about how to prove they wrote their own work.

Paul Beckermann 11:54 How do you avoid that cat-and-mouse game? It could become an endless cycle of chasing students trying to figure out if they're cheating, and that's just a horrible place to live. How do you avoid that?

Amy Onstad 12:09 Right now, I don't, because it's new. But the students I work with I've had for two or three years in a row, so I have the advantage of a pretty good relationship.

I've had conversations with students about it. They come up with solutions for everything — "You could just have us write it in class," or "We'd have to handwrite it." And then, "Well, you could go home and handwrite it if you knew the prompt and just copy from your AI screen." We go back and forth, and they finally said, "Isn't it just on us if we don't know? If we haven't written our own paper all year, won't that show on the AP test?" I said it would — except that scores reflect on whether I'm doing my job effectively. They got it.

It's been a great process to go through with students, because we've had great conversations. But I'm still worried for them moving forward, because I don't know how they can prove they did their own work when they really did.

Paul Beckermann 13:23 We had a guest on the show from California who developed an extension called AI Trust You, which was really about enhancing communication between teacher and student. I know that you and your district have done some things in that area as well. How do you communicate your expectations around AI use to your students, and how do they respond?

Amy Onstad 13:48 They're fantastic — I have the greatest kids ever. Last summer, I was part of a focus group in the district where we made templates for inquiry projects, writing, and researching, using a sort of red/green system — "Can use AI for this skill or not." We also added a "Use with Caution" category, which I worked on with Pam.

For each paper, I have a grid with those templates. I put it up in the progress document and say, "Let's start from the beginning — can we use AI to brainstorm ideas?" Then we have a conversation: yay, nay, talk about why you think we should or shouldn't, and they make their case. By the end, we've come up with what they can and can't use it for, and there's no confusion. They're usually very open to that.

One trend I've noticed is that there's becoming a group that says, "I don't use AI — I just don't, and I'm not going to." And then others who say, "I'll use it for everything." I'm getting subtle vibes that the students who aren't using it feel in some way superior to those who are. It's nothing outward — it's just something I sense knowing them as I do. Does using AI for permitted purposes make someone a bad student? I don't think so. But I also understand the ethics behind why a student might not want to engage with it. It's an interesting little separation.

Paul Beckermann 15:29 That is a very interesting dynamic.

Winston Benjamin 15:33 I appreciate that you've given us a clearer picture of where students are on the use/don't-use spectrum, and you've talked about your expectations around quality of work and your worries about how they'll be judged in the future. Can you give some specifics on the writing process itself — from brainstorming through submission — and what steps you've gone through to help students engage with and navigate AI appropriately?

Amy Onstad 16:30 It depends a little on the context. With literature, we don't use AI much, because we do some test runs — "Let's ask it about this book we just read" — and students quickly see it's inaccurate. We talk about that and move on.

With AP Language, where students are writing synthesis, persuasive, and argumentative essays, I use School AI, a program where I have students enter what they're trying to prove in a very rough form — no crafted thesis or anything. I have them do a three-to-five exchange with the AI about trying to prove something, and I've instructed the AI to push back, to help them solidify what their arguments are going to be.

I've also had students put an introduction into the AI along with the rubric to see what feedback they'd get, and then we discuss that feedback together.

I haven't done a formal full-paper submission before turning in — I have so many students who are choosing not to use AI that I don't want to make it mandatory. It's offered as an option if they'd like to try it.

Winston Benjamin 17:49 That makes sense — providing students options and choice, giving them engagement opportunities.

Rena Clark 17:59 I'm curious whether you've used AI in any other ways — for building relational capacity or getting to know students? I've seen teachers, especially at the beginning of a year or semester, use it to gather information about students, then synthesize that data to help better inform instruction.

Amy Onstad 18:43 Oh, no, I have not done that. That's a good idea, though.

Rena Clark 18:48 It can really help on both the academic and the relational side. Which actually brings me to this — I believe you were part of an AI cohort in your district with Pam. Can you talk about what that was like and some things you learned?

Amy Onstad 19:10 That was great. Anything our digital team does, I'm pretty much going to sign up for. Technology has very much intimidated me for a lot of years — going back to when we went one-to-one in our district and when Schoology came. I did not know how I was going to get through that, other than just going straight forward. Luckily, I had Pam there and Paul too at the time to help us navigate. 

The cohort was a great experience — about 20 of us, which is always a good number. I had a couple of my own colleagues from the English department with me. One of the first things Troy did was put up two images and ask, "Which one's AI and which one's not?" I did not do well. Everybody else was saying, "That's obviously AI," and I was thinking, "Really? What do you mean, 'obviously'?" The image side of AI had not even occurred to me before that moment. Kids are seeing pictures of things and thinking they're real, I'm thinking they're real, and they're not — and people are reacting and overreacting. That was a bit of a wake-up call.

From there, we got into using tools like NotebookLM and Gemini and Gems, and were given time to navigate and test those out. I was in a state of mild panic after that first session, but it got better.

Rena Clark 21:00 The images just keep getting better. I can't tell anymore either.

Paul Beckermann 21:07 Did you run into a favorite tool during the cohort — something that lit the light bulb?

Amy Onstad 21:15 The light bulb was always lighting up, but then it would dim — I'd go back and think, "What was that again?" and run over to my neighbor asking what the thing was.

I like NotebookLM, though I'm still figuring out whether my subject area benefits from it most. I feel like classes that involve a lot of note-taking would benefit more from compiling and analyzing those notes. I do like the Gems, though. I use them to create a kind of guided AI experience — not just open ChatGPT where everything is available, but a space where students put in their planning guides for a content creator assignment and the AI gives them feedback. I set the parameters, which I like. And since it's within our Schoology system, my students who are hesitant about AI don't see it as an extra, separate thing — it feels more integrated.

Paul Beckermann 22:57 So essentially, you write the prompt and share it with students, guiding the learning experience.

Amy Onstad 23:05 Yes, and they have to put in their own documents, so it analyzes their own work, which I also like.

Paul Beckermann 23:12 Very cool. You mentioned that you teach mostly AP and college-bound students, but also different grade levels. Do you see differences in how students across grades approach AI?

Amy Onstad 23:31 Not really. Maybe my seniors use it more practically — that's the letter-of-recommendation fellow. They see it a bit more as a real-world tool. My 10th graders tend to be a little more nervous — "I don't want to use it because I'm scared I'll get in trouble." I don't blame them, because there isn't enough consistency across teachers in how AI is handled and responded to. But there isn't a huge difference by grade — in every grade there are those who are scared and those who are embracing it.

Paul Beckermann 24:09 How about colleagues in other curriculum areas or grade levels, outside the college-bound track you tend to teach? Have you heard different impressions?

Amy Onstad 24:24 I don't talk a lot outside my discipline about it specifically, other than those who were in the cohort. But there's a lot of talk about conversations with students when they do get caught using AI — how those conversations go. It's very different for different teachers. Some kids will say, "I didn't use it," and the teacher says, "I think you did, because I don't think you wrote this poem." And students have told me, "Why is the default always that we did it? Why isn't it a conversation first?"

That's hard to hear, because I don't want to step on that teacher's toes — there's a reason they reacted the way they did — but the student side of it is the frustrating part. Students already feel under scrutiny so much of the time, and this is just another thing that can cause a lot of angst and damage relationships between students and teachers.

Paul Beckermann 25:38 One follow-up: we've talked about within your school, but you also teach at St. Mary's as an adjunct professor working with education students or graduate-level students. What are you hearing from them? Are they talking about the same kinds of things?

Amy Onstad 26:07 The papers they write for me are fairly personal, so it would be hard to use AI for them in a meaningful way. The program I use doesn't check for AI, and I don't sense that they have been.

I do have a colleague who is in a master's program himself, and he said that before he turns in a paper, he has to run it through his university's sponsored AI tool — that's required before submission. I don't know the purpose behind that, but it blew my mind.

As for the content we cover, AI doesn't come up a lot. Many of the students in that class are elementary teachers, so it may just not be as central to what they're engaging with in their classrooms. But it's something I'm now more attuned to.

Winston Benjamin 27:08 You said something I'm still processing — this idea that the default assumption is that students are using AI. Thinking about that, what advice would you have for teachers who haven't yet engaged with AI, or who default to suspicion? What would you say to both groups?

Amy Onstad 27:52 First: you have to get yourself educated. If all you know is "it says 100% AI" — slow down. However you want to learn, do that. We have a fantastic digital learning team and good tech mentors in every building in our district, and I'd encourage people to use those resources.

You have to know that Turnitin — I love it, but it is not accurate for AI detection. You have to know some of the other options. You have to know your students. You have to know tools like Brisk. And you have to talk to the students. You have to talk to the students. You cannot just send an email or write on a paper, "This is showing AI — redo." You have to have a conversation.

And you have to check your body language. Even if you're saying, "We'll talk about it," if your face is saying, "I know you did it," students will pick up on that immediately. When I call a student in, I write: "This is showing up — let's have a conversation." They come in, we sit down, and I say, "So tell me. Here are the paragraphs that are flagging. Walk me through it."

In about 99% of my interactions, students will say something like, "Okay, the hook — yeah, I did use AI for that. Here's what I had before, I just needed to polish it." And then for the next section: "I didn't use it on that." And sometimes they'll say, "Did you use Grammarly Premium?" "I did." Because Grammarly Premium rewrites sentences, and students don't always know that's going to flag. Now they know — and that's a win. They'll say, "Yeah, I used it there — I'll redo that part."

You have to know your tools, you have to know your kids, and you have to give them the benefit of the doubt. And each of those conversations is a chance to teach.

For new teachers specifically, I'm hoping education programs are doing what they need to do to prepare them. College students know about AI personally, but approaching it from the instructor side is going to be a shift.

Winston Benjamin 30:51 Absolutely. It's going to be valuable to consider our relationships with students when navigating AI. And now that you've mentioned all those tools that support teachers in identifying AI use, it's time for our toolkit.

Transition Music with Rena's Children 31:13 Check it out. Check it out. Check it out. What's in the toolkit? Check it out.

Winston Benjamin 31:22 What's in your toolkit? What are you thinking about adding to your repertoire? Paul, Rena?

Paul Beckermann 31:32 Amy got me thinking about NotebookLM. I love it because you have control over the content the AI is engaging with. For example, if I'm doing a Shakespeare unit, I could choose all the resources I want about the Shakespearean era, put them into NotebookLM, then click a shareable link so students can engage with the learning assets I've created — studying based on what I've curated. Or I could put in the whole text of Romeo and Juliet and have students ask questions or find connections within it.

You do have to ask yourself: are they using this in a way that short-circuits the legitimate learning outcomes I have for these students, or is it ancillary to those outcomes? Those are important questions. But I do like NotebookLM.

Rena Clark 32:37 Similarly, Amy talked about Gems, and I find that whether it's School AI spaces, Gems, or Magic Student, using those tools to help build relational capacity and provide another platform for students to ask questions — or for teachers to get to know students — is really helpful. You can gather information, then use it to better understand the dynamics of the class and shape what you're doing to be more engaging. And the nice thing is you're in control and can set the parameters.

I've mentioned this across contexts — from helping a PE teacher build baseball IQ, to analyzing a text, to helping students brainstorm which ancient civilization they might want to study. It really helps, especially in middle school where simply getting started is often the hardest part.

Winston Benjamin 33:47 So true. My tool is going to come from Amy's advice: check out our past episodes of Unpacking Education where we talk about AI in the classroom. And as you do, take Amy's suggestion: pick one tool, dig into that tool, and don't keep trying every new one that comes along. You'll get lost in the options and won't be able to do anything well. Pick one thing, learn it, and try it.

Amy, would you like to throw something into the toolkit today?

Amy Onstad 34:30 I like what Rena mentioned about personalization. I have a lot of students who are beginning to reflect on their high school careers, and I have them fill out a form for their letter of recommendation. But imagine having them fill that out and put it into an AI and say, "What do you see about this person?" — to give themselves a little boost of confidence by having AI look at their awards, activities, jobs, volunteer work, and courses. I think that would be really interesting.

But if I had to leave with one thing: whether students like to use it or not, they are going to be in an AI world. The jobs are going to have AI. Even though it's not yet a standard, we need them to be ready to tackle it — for college, for careers. To withhold it and make it a source of angst and anxiety, to make it an "I caught you" situation, would do them a terrible disservice moving forward.

Rena Clark 35:37 That actually brings us to our next segment — One Thing.

Transition Music 35:41 It's time for that one thing, that one thing, that one thing.

Rena Clark 35:49 We're going to share that one thing we're still thinking about or really want to leave you thinking about. Winston, Paul?

Paul Beckermann 36:04 I am just marveling at how much Amy cares about her students. That is so clear as I listen to her answers and hear how central relationships are to her teaching. She doesn't want it to be a gotcha situation. When she does have to sit down with a student to work through something, she turns it into a learning experience — and that can actually strengthen the relationship in the long run, when it's done with empathy like that. I encourage people to follow that model. Have those relationships and have those conversations with kids. It will go a lot farther than trying to police AI use.

Winston Benjamin 36:52 So true. I'm thinking about what Amy said about the range of students — those who want to use AI and those who don't — and being cognizant of that as you design activities. Can students still be successful with or without AI? Be aware of providing students choice and options, because some may not be comfortable using AI as a tool.

Rena Clark 37:27 I keep coming back to the idea of culture and calibration. I worked in a lot of different high schools, and Amy, I saw the exact same problem — students were confused. What was acceptable in one class was not acceptable in another. I had multilingual students who were accused based on an AI detection tool, never spoken to, and given detention. What does that say about your school's culture?

There's also a calibration issue: schools need to be having these conversations, because when everyone does things differently, you're sending different messages. As a parent — and I happen to have twins who are in the same class — when the flexibility and learning experience one child gets is very different from the other's, and one is being talked to and learning while the other is being punished and experiencing something very negative, that's a problem. How do we calibrate and create a positive culture for students?

Amy Onstad 38:49 I do have one more thing. For teachers who are hesitant or nervous: just remember a couple of things. I graduated in 1987, and I was in middle school when the debate over calculators erupted. "Can they use a calculator in math? Oh my gosh — they'll never learn math!" We got past that. Then Wikipedia came around: "Oh my gosh, we can't let them use Wikipedia — the horror!" Now Wikipedia cites its sources. We have been through this before.

I remember learning about email and being so frightened about it I swore I'd never use it. We've navigated every new thing in education — Chromebooks, learning management systems, all of it. It's going to be fine. We will traverse AI. We will keep learning about it, keep educating each other, and rely on the teammates and experts around us every day.

Paul Beckermann 39:57 We appreciate you joining us today, Amy, and sharing your experiences with AI. Thank you so much.

Amy Onstad Thanks for having me.

Rena Clark 40:07 Thanks for listening to Unpacking Education.

Winston Benjamin 40:10 We invite you to visit us at avidopenaccess.org, where you can discover resources to support student agency and academic tenacity to create a classroom for future-ready learners.

Paul Beckermann 40:23 We'll be back next Wednesday for a fresh episode of Unpacking Education, and

Rena Clark 40:28 remember — go forth and be awesome.

Winston Benjamin 40:32 Thank you for all you do.

Paul Beckermann 40:34 You make a difference.