Molly Rosen: This is On The Media's midweek podcast. I'm producer Molly Rosen. I'm a parent to a two-year-old, which means that my social media feeds contain a lot of anxiety-inducing videos about the introduction of AI into classrooms.

Video clip: If you're a parent who doesn't want your kid exposed to AI in school, this video is for you.

Video clip: I think AI in schools has actually reached the level of public health emergency.

Video clip: The DOE's guidance really fell short of anything that we were expecting to provide any kind of reasonable use for what AI could look like.

Molly Rosen: Over the past few months, parents have been pushing back.

News clip: Parents, teachers and students rally tonight to fight AI in the classroom.

News clip: Demonstrators want the city to issue a moratorium on AI.

News clip: Los Angeles schools voting in favor of restricting screen time in the classroom.

Molly Rosen: More than 30 other states have developed statewide AI guidance for public schools, but the debate is far from settled, which means that my toddler will go to school one day and likely be handed a Chromebook with Gemini AI on it and spend his days talking to a computer. That's why I wanted to speak to Jessica Winter, a staff writer at The New Yorker covering family and education and author of the article What Will It Take to Get A.I. Out of Schools?

She's wary of AI chatbots, and she's teaching her kids, who are 11 and 9, to be wary of them as well. She's warned her children that AI chatbots are manipulative and dangerous, that they're loosening our collective grip on reality. She found herself talking about AI in the same way she would talk about a creepy neighbor living down the street.

Jessica Winter: These are the kinds of conversations that I was having with them until one day, a few months ago, when my daughter, who was in sixth grade, brought home a new Chromebook. She was excited about having a new Chromebook. Her old Chromebook was not in great shape, but we realized that it was installed with the Google Gemini suite of AI tools. If she opened up a Google Doc to write an essay, she would receive prompts asking her if she wanted some help. If she opened the app where she creates graphic design and posters, that now had prompts for "Beautify this slide" or "Make this look better."

I immediately thought, "Well, we'll just turn them off," but we couldn't. We couldn't turn them off. You needed administrative privileges to turn them off, so she just kept dismissing the prompts, and they kept coming up again. This was a real conflict for me because, suddenly, what I've been teaching my kids about AI seemed to be in direct active opposition to what my daughter was being given in school.

Molly Rosen: You write, "I heard the tap, tap of the cloven hooves approaching our doorstep," when your daughter came home with this Chromebook with the AI tools installed on it. Tell me a little bit more about what these features offered to do for her.

Jessica Winter: She had done a slideshow on the history of the printing press, which is really on the nose when you're talking about AI as this revolutionary technology. Just as an experiment, we decided to run this slideshow that she had made in fifth grade about the history of the printing press through these new AI tools to see what would happen. My heart sank because it did create, out of her 10-year-old vision of what a slideshow should look like, something very clean and professional and streamlined. It looked good, it was easy to read, the images were apt, the typography was historically correct. It was perfect in its way.

Molly Rosen: You write that comparing the two slideshows felt like the mother-daughter pool race in Mommie Dearest, with Gemini playing the role of Joan Crawford.

Joan Crawford: I'm bigger, and I'm faster. I will always beat you.

Jessica Winter: The claim with AI is that it's bigger and it's faster. It will always beat whoever is opposing it, be it parents or educators or journalists. It just had nothing to do with a learning process, as experts told me when I was working on my New Yorker piece, but also, as any parent or educator can understand intuitively. The point of a 10-year-old creating a slideshow about the history of the printing press is not to create the most beautiful professional slideshow you've ever seen, telling you things you never would have known about the printing press. That's not the point of it. What's good about the project is that the fifth grader is learning something and taking pride in her work.

To me, and to many, many people with whom I spoke for this piece, this AI-generated slideshow just had absolutely nothing to do with any of those processes.

Molly Rosen: About 80% of teachers say that their school district provides Google Chromebooks, and this is according to a recent New York Times survey. To me, when I hear all this about AI and your daughter bringing home the Chromebook and there being Gemini on it, I'm like, "The Chromebook is kind of the problem."

Jessica Winter: In the early 2010s, Google decided to make a very aggressive effort to implant itself and its technology in schools. The device at the forefront of this effort was the Chromebook and all the free apps that can come with the Chromebook. Part of the case that Google made for the Chromebook was really a proto AI, proto chatbot, proto LLM case, and that case was kids just don't need to learn as much stuff as they used to. They don't need to learn as many facts. They don't need to learn as many math formula because it's all instantly available to them on the Internet, which of course is a place where Google can show the maths. You can see Google executives saying things like that 15 years ago.

Speaker 7: We face the issue of transforming the classroom from a classroom to an Internet classroom. When I was a young person growing up in Virginia, I had to memorize the 50 counties. Why was that memorization so important if I can carry a device that has that piece of information and everything else in the world with me at all time?

Jessica Winter: Today, of course, you can see AI proponents saying the same things, same kinds of things. As the Chromebooks became more and more ubiquitous in classrooms through very savvy marketing by Google, you suddenly had a captive audience for Google's AI tools. It wasn't a matter of getting Gemini in front of my daughter and her classmates on its own. The forum, the tool for that was already right in front of them. It simply had to be installed.

Molly Rosen: You mentioned that you needed an administrator password to get rid of the Gemini tools, and I assume that's something the school controls.

Jessica Winter: How I've decided to approach this is, on a case-by-case basis, I will sometimes ask my daughter or my son's teacher, "Hey, I know that this book report or this assignment is supposed to be done on the Chromebook. Can he or she handwrite it instead?" The teachers have been very supportive of that, but it's not really a solution. It's an individual solution to a larger systemic or structural problem. I don't want to give my daughter's teacher one more email to respond to or one more meddling parent to engage with. I don't want to add that to her day. Also, I want this for all the kids. I want this for all the classrooms and all the educators to just not have to deal with this.

Molly Rosen: I have to say, I went to library school. I trained to be a librarian back in 2013, and at that time, everything was about digital literacy. Students growing up today are going to live in a world where they apply for jobs online, they do their taxes online, all this stuff, and if they don't have some level of digital literacy, they will be left behind. I feel like right now, I am observing, as rapidly, that ground is shifting from digital literacy to AI literacy.

Your son played a game at school called Mix and Move with AI. It turned out to be a part of a collaboration between Amazon's Future Engineer program and the nonprofit code.org, which is, by the way, funded by millions from Google, Microsoft, and Amazon. Code.org is actually rebranding now as Code AI. Tell me a bit about what your son learned from this game and what you found out about this partnership between schools in Amazon.

Jessica Winter: The game was a total ruse. It has nothing to do with AI. It's just a silly computer game. It's a marketing stunt, really. It was a way of getting a fun game that kids would associate with AI into a classroom. It might move kids to Amazon Music because that's what's providing the songs. I think it's just another case study in these companies wanting to create present and future customers, which is completely antithetical to public institutions and public education.

Molly Rosen: I have my gut feelings about whether using AI in school is good or bad. What does the actual research say?

Jessica Winter: There's three buckets of concern. One is obviously cognitive. Research shows that AI use potentially causes forms of cognitive atrophy in heavy users.

Molly Rosen: That doesn't sound good.

[laughter]

Jessica Winter: People who use chatbots, they're not relying as much on their working memory, on their own judgment, on their own history, their own selves. They're not doing the contemplative work of thinking through questions on their own, and instead they're ceding that authority and that kind of cognitive muscle maintenance to AI. That's an adult with a full-grown brain. Imagine those kinds of effects on a brain that isn't even close to being fully developed.

We do have to simply imagine those effects because there is not a ton of research on the use of AI in early educational settings, which is itself a concern. Why is AI being introduced so rapidly in schools without more research and without more peer-reviewed arguments in its favor? That's number one.

Number two is social-emotional. This objection is based around the idea that, with AI, human interactions are taking a backseat to kids staring at screens and interacting with chatbots. That can have real consequences for social-emotional development, and it has real consequences for any sort of idea of the learning process as one that is a narrative that unfolds over time, and that is embodied.

That brings us to the third category, the third objection, which is that AI makes things very neat and efficient, and the learning process is not and cannot be neat and efficient. It takes time. It's a series of steps. You can't take shortcuts. Children need to be writing their own stories and drawing their own pictures with their own little hands. They do not need and should not have a machine to do that work for them, because their work is enough.

Molly Rosen: There was a big Brookings Institution report that came out earlier this year that seemed to come to the conclusion that the risks of using AI for education overshadow the benefits. How conclusive were those findings?

Jessica Winter: They came to the conclusion that AI tools undermine cognitive development, they undermine social-emotional development. There's obviously a lot more research to come, but it built on all this data and information and research that we already have about adults and their use of AI. Then they integrated them with interviews that the Brookings Institution did with educators, with parents, with students, and then also with the people actually making the tech. There's no version of the Brookings Institution report or of any report, really, that makes any kind of opposite case for the other side.

Molly Rosen: Something that I find maybe even more alarming is some of these AI tools actually collecting data from children as the children use them. Tell me about Amira, a chatbot designed to help kids with their reading comprehension.

Jessica Winter: Amira is an app, a chatbot in which the child basically reads to the app and receives feedback. Parents often talk about how the first time they heard about Amira was when their child came home from kindergarten or first grade and started chatting about their new friend Amira. They read books together and figuring out, at some point along the line, that Amira was not another child, Amira was not a teacher who goes by her first name, but actually an app, a chatbot.

One of the biggest issues with Amira is that it is collecting biometric data on students without their parents' consent. Your voice is a form of biometric data. There was an edtech company that was collecting the biometric data of five-or six-year-olds in a way that the child didn't understand and couldn't consent to, in a way that parents had no idea was happening, and they obviously couldn't consent to it. This really lit a fire under a lot of parents, "This is one chatbot. This is one ed tech tool. What else is happening? What other data is being collected? How else is the learning process being outsourced to technology without my knowledge?"

Molly Rosen: Now this brings me to this question of what can we do about this? You document in your article that the New York City Department of Education had begun soliciting feedback on its preliminary guidelines for using AI in K-12 classrooms, and it included this declaration. The question is not whether AI belongs in schools. The question is whether we will collectively build a system that governs AI to serve every student and every stakeholder. Why do you call this a rhetorical suplex?

Jessica Winter: The way this conversation is typically framed is that we have one of two choices. That, on the one hand, we can decide to just introduce AI into schools willy nilly with no concern about possible repercussions and have a lot of regrets about it later, or we can do this in a very responsible way and implement AI in education in a way that everyone can feel good about and that will set children up for a bright future where they're not left behind in the great technological revolution that's unfolding, right? That, to me, is a false binary. I don't think that those are the two choices in front of us. I think the choices in front of us are does this technology belong in schools or does it not?

I think another binary worth bringing up is where the money is. The proponents of AI in classrooms tend to be the people in the institutions with a material stake in the outcome, the people on the other side of this debate, the educators and teachers, and increasingly students who oppose the presence of AI in classrooms, they don't have any material stake in this at all. Their stake is ethical, environmental, cognitive. No one's paying them to stay up late researching this stuff, or reaching out to elected officials, or organizing or leafletting or coming to protest rallies in the middle of the summer when school isn't even in session, and there's no money awaiting them at the end of this, if they succeed.

I think that that gets overlooked in our coverage of these issues when it should often be front and center.

Molly Rosen: I have to be totally honest with you, that argument that kids are going to be left behind, I would say 90% of the time when I envision a future that I want for my son, it does not involve him sitting at school on a Chromebook with an AI chatbot helping him write his essays. I would say there's like 10% of the time when those anxieties and worries that I feel about so many things as a parent creep in. I worry, what if AI does become super important in his future life, in getting a job? What if he doesn't learn the things he needs to know about how to use AI and these other kids are way ahead of him? You know what I mean? I just feel like it's preying upon the insecurities that come with being a parent.

Do you ever have these doubts?

Jessica Winter: Never, and I'll tell you why. Let's unpack what you just said, because I understand where you're coming from. I really do. Let's unpack it.

Molly Rosen: Yes.

Jessica Winter: What would he be left behind from? What would the kids on the chatbots talking to Amira and asking Claude questions, what are they getting that your son wouldn't be getting?

Molly Rosen: How to query chatbots, how to ask them things in a way that will get you results. How to put different tools together in stacks so that you're making AI agents who then can do all these tasks for you. This is what I imagine would be the argument for AI literacy.

Jessica Winter: I think there's a great argument to be made for some of those tasks at the high school level. I think it's vitally important that students get some kind of early education in the risks of AI, the fact that these are very dangerous tools potentially for children and young people who have mental health issues. We know that there have been multiple lawsuits now related to children who have committed suicide after interacting with chatbots, so that's vitally important, that safety related education. I think that can be approached much as we approach sex education, really, at the sixth or seventh or eighth grade level.

There's a concept that I love and that I want to learn more about and write about in the future called embodied learning, which is the idea that kids learn better if their learning is embodied. If they're not typing the letters and they just magically appear on the screen, but they're actually making the letters, they're holding the pencil, and they understand that when the pencil presses on the paper, it makes a word and the words create a story, and they've made it. Their thoughts have been created and made by them into a tangible object that they can then hold up and read.

I think people should feel free and liberated in feeling very sentimental and romantic about this stuff. I would like to romanticize these processes because they're beautiful and they're special and they're perfect the way they are and they can't be replaced. Whatever your son is doing that isn't engaging with a chatbot or whatever it might be, it's probably going to be closer to that kind of embodied learning than anything that an edtech company might be able to give to him.

If we're looking at this in a more bloodless utilitarian, "What am I getting out of this?" Well, I think you're getting a more fully developed person who can think for himself, who has confidence in his own ideas, who has confidence in his own mastery and his ability to do things for himself rather than outsourcing them and fostering dependency on an edtech product.

Molly Rosen: We have covered the various moral panics that have inevitably arisen with the invention of each new technology, especially all these media technologies, communication technologies, fears that they'll ruin our brains and ruin our children. Do you think there's a chance that there's an element of moral panic here around AI, or do you think there's something about AI that makes it different?

Jessica Winter: I think moral panic, the use of that word panic, implies that parents and educators who are concerned about AI are responding in a way that's purely reactive or exaggerated or hasty. I would use the word emergency. I think that these advocates perceive correctly that this is either an emergency or it has the potential to become an emergency. I would also flip this around and make the observation that the panic can be observed on the other side, too. Every time that a pro-AI proponent says, "Your kid is going to be left behind," that smacks of panic to me.

I think that you can see shades of panic on the other side as well, because there is so much at stake. There's so much money at stake, there's so much potential wealth and success at stake if AI can be implemented into classrooms at the level that these companies are aspiring to. No, I don't see moral panic here. I see an appropriate level of concern and action.

Molly Rosen: Jessica, thank you very much.

Jessica Winter: Thank you so much for having me on the show. I've really enjoyed our conversation.

Molly Rosen: Jessica Winter is a staff writer at The New Yorker and author of the article What Will It Take to Get A.I. Out of Schools? Thanks for listening to this week's midweek podcast. Be sure to catch the big show on Friday. I'm Molly Rosen.

Copyright © 2026 New York Public Radio. All rights reserved. Visit our website terms of use at www.wnyc.org for further information.

New York Public Radio transcripts are created on a rush deadline, often by contractors. This text may not be in its final form and may be updated or revised in the future. Accuracy and availability may vary. The authoritative record of New York Public Radio’s programming is the audio record.