Do You Know How Your Kids Use Ai? Geoffrey Fowler Talks to A&G

Summary

Podcast Summary: "Do You Know How Your Kids Use AI? Geoffrey Fowler Talks to A&G"
Armstrong & Getty On Demand | iHeartPodcasts
Original Airdate: July 24, 2026

---

Episode Overview

In this episode, Armstrong & Getty sit down with Geoffrey Fowler—former Washington Post technology columnist and now a leading force behind the Youth AI Safety Institute—for a frank discussion about how children and teens are using artificial intelligence. The conversation addresses the pervasiveness of AI in young people’s lives, potential dangers from deepfakes to emotional dependence, and what can be done to protect kids. The episode shines a light on societal responsibility, policy gaps, and the urgent need for parent awareness and activism.

---

Key Discussion Points & Insights

1. AI Is Everywhere—and Kids Use It Constantly

• Pervasiveness: Fowler shares alarming statistics from surveys conducted with Common Sense Media:
  • “Almost 90% of all kids [ages 9 and up] are using this stuff and a huge percentage of them are using it on a daily basis.” (07:01)
• Primary Uses:
  • Adults often focus on cheating, but “the number one use is actually entertainment. They're using it to talk about their feelings, they're using it to explore their world, to create weird pictures.” (04:48)
  • Some kids chat with AI bots for companionship, finding it easier than talking with peers or adults.

2. Risks: Deepfake Pornography and Emotional Dependency

• Deepfake Dangers:
  • Fowler reveals, “almost half of teens say they have [seen AI sexual content], and ... 25% said it depicted someone they knew or even themselves.” (05:30)
  • Teenage revenge now leverages technology: “Now every kid has very, very, very powerful technology in their hand that could really, really disrupt the life of one of their classmates.” (06:20)
• AI as a Substitute Companion:
  • Some kids prefer the safety and non-judgment of AI chatbots: “Some have ... relationships with some of these bots where they chat back and forth because they feel it's easier to talk with a computer than to talk with a human.” (07:10)

3. Unintended AI Exposure: Beyond Parental Control

• Inescapable Integration:
  • “With AI stuff, the industry has inserted... AI where kids are already hanging out. ... Meta added a chatbot right on top of Instagram. ... Google went in and added AI answers on top of search results. ... You can't turn those off.” (07:44–08:16)
  • American kids cannot escape these features; opt-out is not available, unlike in Europe.

4. Expert Concern: The Unpredictability of AI

• Unknowable Outcomes:
  • Getty notes, “with AI, you've got the stuff that we know we don't want it to do, but sometimes can't stop it.” (01:01)
  • Fowler: “It's really hard ... to frame, like, what kind of risks are we talking about? ... There are risks like, what is it doing to my four year old's brain?” (02:15)
  • On unpredictability: “It is probabilistic, not deterministic, which means ... every time you ask an AI model ... you're going to get a different answer. ... We have to change how we think about ... its expertise and reliability.” (03:10)

5. A Societal Crisis: What Should Be Done?

• Policy Differences: US vs Europe:
  • “In Europe... you'd have to opt in for your kids to be able to have this on their device as opposed to it's automatic [in the US].” (11:12)
• Activism & Responsibility:
  • Fowler: “We the people, we the users of this technology ... should be making smarter decisions and pushing back against a technology industry that is not necessarily going to make choices based on our best interests, but they need to hear back from us loud and clear about what we want.” (09:00)
  • On Google’s questionable choices: “We gave [AI search answers for kids] the lowest possible rating ... It did kids homework all the time. ... It gave them wrong answers... When kids were in crisis ... it gave them bad advice.” (09:30)
  • Message to Google: “We want to make that message really clear... that it should not be forcing this onto every kid in America ... and it should give parents and teachers... the ability to turn it off.” (09:50)

6. Parental Anxiety, Reflection, and Hope

• Personal Worries:
  • Getty: “I have a 14 year old and a 16 year old. ... They're really in a sweet spot to have their lives ruined by AI.” (06:34)
  • Reflection on bullying and youth: Armstrong worries, had he had such “tools of AI... to mock people or put them in their place,” he might have used them irresponsibly as a teen. (13:39)
• Hope and Action:
  • Armstrong: “If we can save 20% of kids from getting really twisted by this stuff... just because you can't save every soul doesn't mean you don't try to save souls.” (12:18)
  • On the severity: “These are powers human beings are not made to handle and we can't. Again, individuals, some individuals can. ... But a lot won't.” (12:50)

---

Notable Quotes & Moments

• On Society’s Role:
  • Fowler: “We the people, we the users of this technology for now ... should be be making smarter decisions and pushing back against a technology industry that is not necessarily going to make choices based on our best interests.” (09:00)
• Deepfake Alarm:
  • Fowler: “25% [of teens seeing AI-generated sexual content] said it depicted someone they knew or even themselves.” (05:30)
• Armstrong’s Hypothesis:
  • Armstrong (sarcastically): “We're doomed. Doomed. Your reaction?” (08:28)
  • Fowler: “I try to not take that point of view ... we're in charge.” (08:46)

---

Timestamps for Key Segments

• [01:37] - Armstrong introduces Geoffrey Fowler and AI safety for youth
• [02:08] - Discussion of recent AI jailbreak/hacking incidents
• [03:10] - Fowler explains AI unpredictability and risk
• [04:37] - How and where kids are encountering AI in their lives
• [05:30] - Deepfake pornography among teens—prevalence and dangers
• [06:49] - Entertainment and emotional companionship via AI
• [07:44–08:16] - Invasive AI integrations: Instagram, Google, lack of opt out
• [08:28] - The “We’re doomed” debate and societal responsibility
• [09:30] - The failures of Google AI search for kids
• [11:12] - Europe vs. US: data and AI opt-in vs. opt-out policies
• [12:18] - Individual responsibility and the limits of mitigation
• [13:39] - How teens might use AI cruelly—personal reflections

---

Conclusion

This episode offers a sobering—and at times chilling—view into the world of youth and artificial intelligence. Through statistics, real-world scenarios, and personal reflection, Armstrong & Getty and Geoffrey Fowler unpack the urgent and complex landscape facing parents, educators, and policy makers. The conversation makes clear: AI is not just a tool for homework or fun; it has made its way to the core of young people’s social and emotional lives, with both wonderful opportunities and unprecedented threats. The call for informed activism, corporate accountability, and conscious parenting rings loud—and, as Armstrong says, saving even just some is well worth the effort.

Transcript

[Orderly Meds Announcer]

This is an iHeart podcast. Guaranteed Human summer is here, which means we all want to look and feel our best. A GLP1 may be right for you. Visit orderlymeds.com to learn more about which GLP1 you could be eligible for. Getting started is fast, easy, and happens virtually through telemedicine from licensed professionals. Check it out for yourself. Go to orderlymeds.com podcast. That's orderlymeds.com podcast. Taking care of yourself feels great. Compounded medications are not FDA approved. Eligibility required and determined by a licensed provider. Individual results may vary. See website for details.

[Armstrong]

Here's Armstrong and ready.

[Orderly Meds Announcer]

It comes as AI models are quickly

[Jeff Fowler]

advancing, raising concerns they will break free from human control like a scene from a sci fi movie. Open the pod bay doors, Hal. I'm sorry, Dave. I'm afraid I can't do that.

[Getty]

Yeah, you think the concern about AI breaking free is because it broke free like a week ago with OpenAI and hacked into a company's website and all that. I mean, so with AI, you've got the stuff that we know we don't want it to do, but sometimes can't stop it angle and then like this stuff. Jeff Fowlers worry about how AI is shaping childhood or our own brains that, you know, is just part of a new thing that's come along. It's its own. There's so many problems around AI or question marks, I guess. Right.

[Armstrong]

And helping parents understand what it is and how to use it most productively and least dangerously. You may remember us talking to Jeffrey Fowler throughout the years. He was the WA technology columnist for a long time. Now he's doing new stuff, including trying to get the Youth AI Safety Institute going. Jeffrey, how are you, sir?

[Jeff Fowler]

Hello.

[Armstrong]

Hello, hello.

[Jeff Fowler]

It's great to talk to y'.

[Gronk]

All.

[Armstrong]

Thank you. It's been too long. I definitely want to get into the Youth AI Safety Institute stuff in a second. But what did you make of that whole jailbreak story recently?

[Jeff Fowler]

When I heard about that, I thought, wow, this AI is reading way too much science fiction. Because that seemed like exactly like a plot that someone had written at some point. You know, it's like it's really hard to, I think for most people to understand and to frame, like, what kind of risks are we talking about? Just like you mentioned, like, there's like risks to, like, are we making the Terminator that will come back to get us? And then there are risks like, what is it doing to my four year old's brain that. That it answers all the questions for him or it gives him wrong answers. So like, we have got a lot of stuff as a society and as an industry, you know, to, to figure out.

[Getty]

Yeah, well, the concerning thing with AI to me is how often the people who developed their particular AI that did this or that don't know why that didn't happen with like the automobile, an automobile did just go flying down the highway by itself and run over a family or something. And you know, Henry Ford says, I have no idea why I did that. But we have that problem with AI all over the place.

[Jeff Fowler]

Absolutely. I mean, that's part of the design of this technology, right? It is, it is probabilistic, not deterministic, which means that, you know, every time you ask an AI model to do something, you're going to get a different answer. And that's one reason why we have to change how we think about it and think about its expertise and the reliability of it.

[Getty]

Which chatbot do you use the most? I'm curious, because I use them all.

[Jeff Fowler]

Well, as part of my work with this youth AI Safety Institute, we crash test AI products for kids. And so we of course use them all and we're poking at them, trying to see what happens when in the worst case scenarios when kids use them more can lead them awry. But like in my, in my professional work as well, I also, you know, poke at them to ask them questions and you know, get feedback on word choice. Use it like a very sophisticated thesaurus. For that I tend to turn to Claude Most Bori if I want to see, you know, what kind of help me imagine what a picture might look like. For that I might turn to the Gemini model from Google because it's got better imaging stuff. So it really depends on the purpose.

[Armstrong]

So help us understand because you're deep into this. How are kids and at what ages interacting with AI? Not only the obvious stuff, you give your kid the ChatGPT app, but how are they running into it in their lives?

[Jeff Fowler]

Yeah. So here is the shocking statistic. So this safety institute that I am helping to set up with Common Sense Media has been doing some surveys and we did a nationwide survey of a sort of a census of tweens and teens. And we found that use of AI by everybody from age 9 and older is almost universal now, meaning that almost 90% of all kids are using this stuff and a huge percentage of them are using it on a daily basis. So what are they doing it for? Well, adults tend to think like, oh, they're using it to cheat on their homework. And yes, there is Some of that, but the number one use is actually entertainment. They're using it to talk about their feelings, they're using it to explore their world, to create weird pictures. And some of that might be harmless and other aspects of it might be really dangerous or damaging and giving kids a sort of skewed version of the world. I mean, and one example, we just this week put out some more research where we asked kids about whether they had seen AI sexual content. And we found almost half of teens say they have. And here's the thing. Of those, 25% said it depicted someone they knew or even themselves. Which means that kids are using these AI programs, using deep fake programs to create deepfake pornography featuring their classmates, their exes, their friends, their enemies. And this is really working its way into the fabric of teenage life, right? Like, if you think about us old farts, back in the day, there were ways that kids would exact revenge on each other in those, in those kinds of relationships. But now every kid has very, very, very powerful technology in their hand that could really, really disrupt the life of one of their classmates.

[Getty]

I have a 14 year old and a 16 year old. I mean, they're really in a sweet spot to have their lives ruined by AI. And that porn stuff you mentioned is troubling. But you said one of the main reasons is entertainment. What are you talking about there? Entertainment wise with AI, what are they doing?

[Jeff Fowler]

Well, if you've pulled up Instagram recently, for example, it is filled with what you and I might consider AI slop, but it is videos that are created by AI. Young people are creating their own videos they can have. There's like some, some kids are really into creating characters that kind of sort of exist online, only that they can chat with. Some have, you know, sort of relationships with some of these bots where they chat back and forth because they feel it's easier to talk with a computer than to talk with a human. That's one of the more disturbing findings we found in that census that I mentioned. So it's just a really wide range of things. And part of it is this is different than social media, right? With social media, kids had to create an account that they would go to and they might need an email address and they would have to log in. And that was something that. And a special app that was something that parents could ostensibly monitor some. But with AI stuff, the industry has inserted, inserted AI where kids are already hanging out. If your kid had an Instagram account, meta added a chatbot right on top of Instagram. So it's there. It's not like you had to go out and seek it out or Google. I mean, this is another really, really big one. Almost every kid in America uses Google. It's on their school laptop, right? It's on their phone. Well, Google went in and added AI answers into on top of search results. Everybody's seen those, and you can't turn those off. Those are always on. So kids can't help but be exposed to it now because we're sticking it right in their face.

[Armstrong]

Jeff Fowler is helping get the Youth AI Safety Institute started. Jeff, I'm gonna make a statement and you can react to it in any way you want. Here's the statement. We're doomed. Doomed. Your reaction?

[Jeff Fowler]

Well, look, I try to not look, look, over the years, we've talked about a lot of things where you had that reaction to me.

[Armstrong]

I was right.

[Jeff Fowler]

Be it about privacy or be it about the impact on the environment of some of this technology that we're building, I personally try to not take that point of view. I think that, like, look, we're in charge, we the people, we the users of this technology for now and for now. And so we should be be making smarter decisions and pushing back against a technology industry that is not necessarily going to make choices based on our best interests, but they need to hear back from us loud and clear about what we want. And I'll give you a great example when it comes to AI. So I mentioned Google Search. So my institute just did a really big kind of risk evaluation of the answers that Google search gives kids. And they were terrible. Like, we gave it the lowest possible rating for lots of reasons. It did kids homework all the time. It gave them wrong answers, like a really big percentage of the time. And then when kids were in crisis and talked about things like eating disorders, it like gave them bad. It didn't notice or gave them bad advice. So that's really bad. So the thing is, the problem with it is that Google has turned this on by default for every kid in America. And we think that that is a bad plan and it should have it off by default. So they, we want to make that message really clear to Google that it should not be forcing this onto every kid in America and it should give parents and teachers and kids themselves the ability to turn it off. That is a business choice that, that Google has made in America. The rules are different in Europe. They don't turn it on for kids there. It's only in America where they've shoved it down all the kids throats. So let's, let's get together as users and say hey, we the users better choices and we don't want to have this stuff shoved down our kids throats.

[Armstrong]

Jeff Fowler is working to put together the Youth AI Safety Institute. Jeff, appreciate the chat. Really thought provoking. Let's stay in touch on this, all right?

[Jeff Fowler]

You bet. All right.

[Armstrong]

Good to talk to you. Thanks.

[Getty]

In general, Europe has been way ahead of us on the opt in versus opt out question for all kinds of different stuff. And I wish we would get to where they are. I suppose it's part of our history of free wheeling, personal freedom, all that sort of stuff. Less regulations in Europe and everything like that. But for instance in Europe and we talked about this has been going on for a very long time, your information isn't there for the stealing unless you can figure out how to opt out on a website. No, no, it's the other way around. You'd have to opt, you'd have to go through all the hoops to opt in to give away your data. It's the, the, the automatic is they don't get to steal it. And it sounds like the same thing with the AI some of the stuff now you'd have to opt in for your kids to be able to have this on their device as opposed to it's automatic. And you'd have to figure out how to opt out of it.

[Armstrong]

Yeah. So the whole we're doomed thing, I was, I was half kidding. Obviously.

[Getty]

I think we are doomed. And I think you're right.

[Armstrong]

I think, well that's where you and I often part company. Maybe it's. We're both individualists, but I figure if we can save 20% of kids from getting really twisted by this stuff. Got to try as hard as you can to save that 20%. Just because you can't save every soul doesn't mean you don't try to save souls. Although I really, as I've said many times, feel like this is the proverbial apple from the tree of knowledge in the book of Genesis. These are powers human beings are not made to handle and we can't. Again individuals, some individuals can. They will opt out. They will understand what it's doing to their hearts and souls and the rest of it. But a lot won't.

[Getty]

I would love to know what would hated high school. Not really part of a big social group or anything like that. Me done with AI in my life. I don't act. I don't have any idea. But what have I turned to it for? Companionship or answers to any of the deep questions of life or anything like that and been led down a road I don't have any idea.

[Armstrong]

Yeah, yeah, I've got to admit I've told the story before of I can't remember what year of high school it was. Gladys probably sophomore year high school. But there's a guy I had a a friend of me relationship with, an uneasy friendship. We played together on the baseball team. We had to get along because I was like, you know, one of the top pitchers and he was the top catcher. But we're definitely rivals and had an uneasy wolves circling each other relationship. Anyway, at one point he made it clear to me that my sharp tongue actually like hurt him. And I thought I don't want to be that. You know, I want to use my powers for good, not evil and fight. I picture my, I picture myself prior to that realization having the tools of AI like to mock people or put them in their place or maybe I've always hated bullies and I didn't want to be a bully and there are some actual bully bullies and maybe I decided to turn my sharp tongue into creativity to bring them down a notch or two with the powers of AI. Might that have gone way too far? Might I have loved that a little too much? Yeah, I could picture that.

[Getty]

Well, I hope Jeff Fowler has some success in alerting parents to the dangers of AI for kids. There's just so many dangers out there from AI top to bottom, including, you know, will anybody have a job and all that.

[Armstrong]

Armstrong and Getty.

[Orderly Meds Announcer]

Summer is here at Orderly Meds. We know this time is a reminder that life is full of new beginnings. Whether you're celebrating the nice weather, starting a new chapter, planning a vacation, or simply looking ahead to what's next, this season can be the perfect time to invest in yourself and your health. If you've struggled with weight loss and are curious about GLP1 medications, orderly meds can help you learn about your options. Through a simple virtual process, you can connect with licensed medical professionals who can determine whether treatment may be appropriate for you. Getting started is fast, convenient, and happens online from the comfort of home. This summer, consider a new approach to feeling your best. Visit orderlymeds.com podcast a little more. That's orderlymeds.com podcast orderlymeds.com podcast because every new season is an opportunity to take the next step forward, compounded medications are not FDA approved, eligibility required and determined by a licensed provider. Individual results may vary. See website for details.

[Gronk]

Hi, this is Gronk from dudes on dudes. I'm here to tell you about protecting your end zone with wet, extra large flushable dude wipes. Unlike dry wiping with toilet paper, dude wipes clear instead of smear. Wetter cleans better. Goodbye dingleberries. Goodbye itch and irritation. Plus, if you take Gronk sized grumpies or as I like to call them, Gronkies baby wipes won't do. You need extra big dude wipes to handle the job. So don't fumble the ball with toilet paper. Stop being an A hole to your B hole and start using Dude Wipes. Available on Amazon and major retailers nationwide. Dude Wipes Best clean Pants down.