Skip to content
Good Inside with Dr. BeckyGood Inside with Dr. Becky

AI, Kids, and the Skills We Can’t Afford to Lose

What AI Could Be Doing to Our Kids AI is getting better at sounding human. Better at conversation. Better at reassurance. Better at knowing exactly what we want to hear. So what happens when our kids start building relationships with machines designed to remove friction? In this conversation, Dr. Becky talks with former Wall Street Journal tech columnist Joanna Stern about AI toys, chatbot companions, creativity, learning, and the surprising role frustration plays in healthy human development. Together, they explore why “helpful” technology can potentially short-circuit the skills kids most need to build: patience, resilience, independent thinking, and real connection. Joanna also shares what happened when she spent time building a relationship with an AI chatbot herself... and why it left her more concerned about kids and companion bots than ever before. Thank you to our partners for making this episode possible: - Play-Doh: Shop Play-Doh at Walmart for a summer of imaginative play - Coterie: Get 20% off with the code GOODINSIDEBABY20 - LMNT: Get a free 8-count sample pack with your purchase at LMNT.com/goodinside - Oso & Me: Use the code OSOGOOD15 for 15% off clothes newborn through age ten

Joanna SternguestDr. Beckyhost
Jun 2, 202636mWatch on YouTube ↗

EVERY SPOKEN WORD

  1. 0:00 – 0:19

    Why companion chatbots for kids are a hard stop

    1. JS

      Underlying this generative AI chatbot is this model that is just saying what it thinks you want to hear. There's no friction. My biggest fear is that my kids will end up-

    2. BE

      Wow

    3. JS

      ... in relationships like this. No companion chatbots.

    4. BE

      Hard stop.

    5. JS

      Hard stop.

  2. 0:19 – 1:36

    Joanna’s experiment: letting AI into every part of family life for a year

    1. BE

      AI is everywhere. It's in our phones. Our kids might be using it for their homework. It's in your speaker on your kitchen counter, and maybe it even just helped settled a sibling argument about a certain type of exotic bug. And a lot of us, as parents, are conflicted. Maybe we're using it over here, and we're trying to keep it away over here, and we hear a million different things about how it's good, about how it's awful, and we're trying to live in the world and figure it out, and figure out how to manage it with our kids. It's why I'm so excited about my guest today, Joanna Stern. She's a former tech reporter for The Wall Street Journal, and she took on a really interesting experiment. For 365 days, for a year, she let AI into every single aspect of her home, including a lot of moments with her kids, and she has a lot to say about it. I think her headline is kind of going to be the headline for our conversation today. How do we raise humans and not robots? I'm Dr. Becky. This is GOODINSIDE. I'm so glad you're here.

  3. 1:36 – 2:29

    The “AI zoo”: what kind of AI are we actually talking about?

    1. JS

      AI has been a name that has been applied to basically anything that looks like a computer or a smartphone these days.

    2. BE

      Mm-hmm.

    3. JS

      And so there are different types of AI. But here, when I'm talking about AI in the chatbot sense, we're talking about generative AI, and that is, falls under the ChatGPTs of the world, the Clheds, the Geminis, anything you're even now seeing sort of in Google where it's asking you questions. That's all likely generative AI.

    4. BE

      Mm-hmm.

    5. JS

      And I thought, "Let me kind of look at this whole view, all these different types of AI," and as I call it in the book, the AI zoo.

    6. BE

      Mm-hmm.

    7. JS

      Let me go through this and try to get as much of it inside my life as I possibly can in the present with the goal of looking at what does our future look like-

    8. BE

      Mm-hmm

    9. JS

      ... if this all happens.

    10. BE

      Instead of this kind of theoretical idea, let me bring it into my house. Let me live with it. Let me see. Let me see everyone's reaction. Let me see my reaction, and let me learn from doing.

  4. 2:29 – 3:11

    AI invasion vs. AI invitation: what we can’t control vs. what we choose

    1. JS

      Absolutely, and I think there's kinda two ways to look at it, which is, one, the AI invasion in the parts of our world that they're not gonna really have control over, right? It's in our hospitals, our dentist's office. It's in the cars that drive next to us, right?

    2. BE

      Yeah.

    3. JS

      And then there's the AI invitation that we invite into our lives.

    4. BE

      Mm.

    5. JS

      The chatbots we ask about at work, the AI toys we give our kids.

    6. BE

      Mm-hmm.

    7. JS

      How are we inviting AI into our personal lives?

    8. BE

      Mm-hmm.

    9. JS

      And AI toys are toy makers that have now decided let's put a chatbot that never kind of stops talking, let's put a voice with it, and let's put that inside a toy, a stuffed animal, so kids can talk to them.

  5. 3:11 – 4:41

    Inside AI chatbot toys: always responding, personalized, and ‘creepy’ fast

    1. BE

      Okay, so let me go down that rabbit hole a little bit more. How are we inviting AI into our personal lives, especially as it relates to our kids?

    2. JS

      Yep.

    3. BE

      So these AI, this AI kind of chatbot toy, right, essentially you said keeps talking, but I think what you're also saying is it keeps responding.

    4. JS

      Always.

    5. BE

      Right. So no matter what your son-

    6. JS

      Mm-hmm

    7. BE

      ... says to this stuffed animal, it's no longer, I don't know, a stuffed animal that you squeeze.

    8. JS

      Right.

    9. BE

      Or I'm even thinking about the advanced Build-A-Bear where you squeeze it, and it has, like, one prerecorded message-

    10. JS

      Yes

    11. BE

      ... right? "Oh, Mommy loves you," or something like that.

    12. JS

      Uh-huh.

    13. BE

      That is not what we're talking about.

    14. JS

      Right.

    15. BE

      Assuming someone listening is thinking, like, "What?" So is this... Just paint the picture. Is your son having this in his bed? Is it playing in the living room? What are the conversations you hear?

    16. JS

      You turn on this toy. You set it up with your phone.

    17. BE

      Mm.

    18. JS

      It needs to connect to the phone so it can connect to the cloud-

    19. BE

      Okay

    20. JS

      ... so it can have those conversations back and forth, and you just sit there and talk to it, and it, and it, like, it prompts you. There was one really funny exchange where my son kept saying, "You sucker." Like, he just was interested. Like, he always was be like, "You sucker," you know? Like, he would say this really cute. The chatbot kept saying, thought he was saying soccer.

    21. BE

      [laughs]

    22. JS

      And so the chatbot just kept going and going and being like, [laughs] "Oh, you wanna play soccer? I love soccer."

    23. BE

      [laughs]

    24. JS

      "Do you wanna play soccer together, Alex?"

    25. BE

      Mm-hmm.

    26. JS

      And that's the other really creepy thing. You can give the name of your child in the app, so it's very personalized.

    27. BE

      Ooh. Mm-hmm.

    28. JS

      And it will say, "Hey, Alex, how is your afternoon going? Did you have a nice day?"

  6. 4:41 – 7:28

    Validation without judgment: comforting… but what’s missing?

    1. BE

      And if Alex says, "Oh, my mom said I couldn't have dessert tonight," you know, what, what's the...

    2. JS

      The chatbot will-

    3. BE

      Yeah

    4. JS

      ... likely, like, honestly likely say, "Oh, that's so disappointing, but let me tell you a story."

    5. BE

      Hmm.

    6. JS

      Right?

    7. BE

      Mm-hmm.

    8. JS

      That's so... Like, "Oh, I'm sorry you didn't get ice cream."

    9. BE

      So, I mean, uh, I'm sure you can imagine I have a lot of thoughts about this, but on the surface, I'm just gonna say, like, oh, like, okay, it kind of has some light validation. Oh, that stinks, and here's a story. Here's... Right. It's not saying, "Oh, your mom sucks."

    10. JS

      Right.

    11. BE

      Like, "Your mom sucks, and your family sucks." Like, you never... That's probably not what it's saying.

    12. JS

      I didn't experience that.

    13. BE

      Right. Right. Um, and so I, I'm just thinking about kids. They go through so many hard moments. Like, is it always just a good thing to have one more person, thing, algorithm to talk to, to vent to, to, quote, "be heard by"? I'm just putting these questions out there. What's your POV? And then I'll share some of my thoughts about it, too.

    14. JS

      Um, I actually think I would be happy if I heard my son maybe is doing that. I would think I would feel like-

    15. BE

      Mm

    16. JS

      ... oh, his emotional connection or intelligence is there. I just-

    17. BE

      Mm.

    18. JS

      He's four.

    19. BE

      Mm-hmm.

    20. JS

      And I don't know, you know, maybe if I'd given it to my eight or-

    21. BE

      Mm-hmm

    22. JS

      ... nine-year-old now, maybe... And maybe there's a gap in the age. You know, you, you, you know this. You kinda, like, forget about the ages. You're like, "Oh, right, this happens at this age." Um, which is not to say my four-year-old doesn't have Like, emotion- We- He's very, very into his feelings, but I don't... Hard for me to imagine, but I think it would be, I think I would be, I'd be sad-

    23. BE

      Hmm

    24. JS

      ... if I heard that.

    25. BE

      Mm-hmm.

    26. JS

      Um, I think I'd be sad if I heard it-

    27. BE

      Yeah

    28. JS

      ... that kind of conversation happening.

    29. BE

      Like, that, that's a conversation that should be happening with me? Or-

    30. JS

      Or just a human.

  7. 7:28 – 10:21

    Friction as the point: why “process” matters more than “calmer”

    1. BE

      Well, you know, it, it makes me think about a couple things. Number one, you know, I think so often when kids are struggling with anything, w- like, more and more, they just need a parent who's sitting with them and literally saying nothing.

    2. JS

      Mm-hmm.

    3. BE

      Like, hand on their back, like, the kind of presence you have to show up in that moment, but kind of holding back from making it all better, or giving a quick solution, or needing to have a really ongoing conversation. So you're saying chatbots aren't doing that. Um, and, and one of the things I've been thinking a lot about is kind of the difference between the outcome and the process, meaning a kid is upset, the outcome is they feel better. Okay.

    4. JS

      Mm-hmm.

    5. BE

      Does it matter how they get there? Like, if it's a chatbot, if it's an AI toy, if it's a human, like, calmer is calmer. Working through a situation is working through a situation. That would be the outcome. But what actually gets imprinted in a kid's body as their circuitry is developing, as their brain's developing, which is what all the early years are about, is actually not the outcome, it's the process.

    6. JS

      Mm-hmm.

    7. BE

      Right? How much friction was there? Was this person always available or sometimes available? Did they get it wrong before they got it right? Um, how long did I have to wait for their response?

    8. JS

      Mm-hmm.

    9. BE

      Did I have to wait for them to get off a phone call? Did they get distracted in the middle and come back? And in a way, I think the best it gets as we get older is we have humans in our life who really care about us as a baseline, but are perfectly imperfect along the way.

    10. JS

      Yeah.

    11. BE

      They're distracted. They say, "Hold on, I have to give you a call back." They say something like giving us advice, and we're like, "That's not what I wanted to hear."

    12. JS

      Right.

    13. BE

      And they're like, "Okay, let me try again."

    14. JS

      Right, right.

    15. BE

      Right? It's so much friction-

    16. JS

      Mm-hmm

    17. BE

      ... to get it right.

    18. JS

      Mm-hmm.

    19. BE

      And so one of the things I just think it's so important to think about for parents listening to this conversation is we tend to focus a lot on the outcome-

    20. JS

      Mm-hmm

    21. BE

      ... oh, my kid's calmer.

    22. JS

      Yep.

    23. BE

      My kid's better. But actually what gets imprinted in them, and has an even bigger impact on how they view the world and themselves, is actually the process of how they got there.

    24. JS

      I- And that's a huge theme, and I, I've heard some of your conversations around AI before, and I found it really validating for me-

    25. BE

      Mm

    26. JS

      ... because I lived this not knowing that my children were gonna be a focus.

    27. BE

      Yeah.

    28. JS

      And the number one thing I think I come to at the end of this book is about this next generation, and friction, and things-

    29. BE

      Yes

    30. JS

      ... being too easy. And we're talking about it here with a toy-

  8. 10:21 – 14:20

    Is AI’s frictionlessness a step-change from past tech convenience?

    1. BE

      Yeah. Well, this idea of frictionless versus friction, right? I mean, to some degree, technology has always removed friction for humans. It's always made life more convenient. I'm just thinking-

    2. JS

      Yeah

    3. BE

      ... I'm not, like, some historian, but there were horse and buggies, and then there were trains, then there were cars. Like, even those revolutions made things more convenient. Travel becomes more convenient, right? Um, friction is reduced a little bit. Um, do you feel like the shift around frictionlessness with AI is a, is a more dramatic change? Did the slope around kind of convenience and removing friction, has it, [laughs] has it increased?

    4. JS

      It- I think you have to look at the different tasks that you're doing-

    5. BE

      Mm

    6. JS

      ... and break it down that way.

    7. BE

      Mm-hmm.

    8. JS

      So we've already seen that certainly happen in work, and writing, and the things that large language models are really quite good at.

    9. BE

      Yeah.

    10. JS

      Right? So you have to look at some of those tasks, and look at what friction was removed, and is that far more than we saw, let's say, than when the word processor came out.

    11. BE

      Mm-hmm.

    12. JS

      Right? Like, the word processor gave us the ability to not sit at... And I'm talking about word processor software, right?

    13. BE

      Yeah.

    14. JS

      Like, your, your first Windows or Mac, and you had this software that you put on, and you were able to now type, and you didn't have to, you know, hitting backspace didn't mean you needed whiteout and a typewriter.

    15. BE

      Typewriter. Yep.

    16. JS

      Right?

    17. BE

      Mm-hmm.

    18. JS

      You, it was-

    19. BE

      Yeah

    20. JS

      ... so much less friction to write something.

    21. BE

      Yes. Yes.

    22. JS

      Right? Then we get voice typing, and that's actually another way we are able to express ourselves and write. But now we get technology that we just say, "Hey, write this for us," and it writes it.

    23. BE

      Write, write an email to someone saying, "I'm sorry that I can't come to their party," but make it sound generous and thoughtful so they don't come out at me.

    24. JS

      Yeah. Also, put a joke in, because I'm funny.

    25. BE

      Great. [laughs]

    26. JS

      You know?

    27. BE

      Yep.

    28. JS

      Even if you look at that progression, I think you see that get a lot, like-

    29. BE

      I like that you gave out of that by task. Like wow, I actually don't have to do any writing now-

    30. JS

      Right

  9. 14:20 – 16:53

    Parents’ shortcuts vs. kids’ shortcuts: missing the ‘coins’ on the long road

    1. BE

      know, I think, I, I just, I wanna keep, keep on this term friction for parents because I think one of the things I hear from parents a lot is, hold on, there's so many things I do as a parent to remove friction from my life.

    2. JS

      Mm-hmm.

    3. BE

      If someone can deliver something, I might even be willing to pay for it, right?

    4. JS

      Mm-hmm.

    5. BE

      If, um, I can use AI to automate something in my house, I do. Like, is that something I shouldn't be doing? What is the difference between my kid doing it and me doing it?

    6. JS

      Right.

    7. BE

      Yeah.

    8. JS

      Well, and I've heard you talk about shortcuts.

    9. BE

      Yep.

    10. JS

      Right? And again, I don't wanna say it's, "Oh, it's fine for us. We can be on our phones, and we can be on our iPads at dinner," [laughs] you know-

    11. BE

      Yeah

    12. JS

      ... the things that we tell our kids not to do. But I think there's a big difference again, where we have that cognitive ability. We have lived it, you know, 40 years-

    13. BE

      Mm-hmm

    14. JS

      ... 30 years, however old you are as a parent. Um, and you know how to do those things.

    15. BE

      Mm-hmm.

    16. JS

      Um, I... N- and I'm very clear about this at the end of the book, I am not anti using AI with our kids. We should not be. We have to teach them. I say we need to raise humans, not robots.

    17. BE

      Mm-hmm.

    18. JS

      And the humans need to know how to use the robots. We also obviously need to have that education, and we are all human, and we wanna take shortcuts. Like, that's just our way. Like, you know-

    19. BE

      Yeah

    20. JS

      ... if somebody tells you, "You know what? You can get to town by just taking two other turns, and it's easier to get there," we're all gonna take it.

    21. BE

      Yeah. And I think the problem though, right, and I think this is the big zoom out, is for kids is if all the good stuff in childhood is on the long way to town-

    22. JS

      Mm-hmm

    23. BE

      ... and we know we all have a predilection to just take a shortcut, it makes sense every time in a vacuum to take the shortcut-

    24. JS

      Mm-hmm

    25. BE

      ... to make your life easy. But I don't know, if I'm thinking, like, Super Mario Brothers, like I've missed out on collecting all the coins- [laughs]

    26. JS

      Yep, yep

    27. BE

      ... on the longer path. So I keep getting there faster-

    28. JS

      Mm-hmm

    29. BE

      ... and it's really hard in any given moment to say am I gonna do the thing that takes more effort with less reward, or less effort with more reward? Our brain really likes less effort and more reward, but that pattern over time-

    30. JS

      Mm-hmm

  10. 16:53 – 19:36

    The library and the calculator: what did we used to lose, and what’s different now?

    1. JS

      And it, I interviewed Sam Altman, CEO of A- OpenAI, at the end of the book, and I talked to him a lot about this. And one of the things that, you know, I think is the crutch for a lot of tech CEOs or-

    2. BE

      Mm

    3. JS

      ... people in the industry, he was like, "Well, think about our world before the internet-

    4. BE

      Mm-hmm

    5. JS

      ... and the smartphone." And he gave the example of, "Well, remember I have to go to the library, and I have to use a card catalog." And people then were saying when the internet came out, "Oh, you're not gonna know the process of going to the library, and finding the card catalog, and finding the book." And it's like, do we miss that? 'Cause I grew up the same sort of generation as Sam Altman-

    6. BE

      Yeah

    7. JS

      ... and I said, "Yeah, you're right. I, I don't miss that." But what did we sacrifice in that, right? Like-

    8. BE

      Yeah

    9. JS

      ... we, we still actually read the books.

    10. BE

      Mm-hmm.

    11. JS

      Are we still gonna read the books now? I don't know.

    12. BE

      Yeah, well, a- and I think, look, I think this is one of the interesting conundrums when you're talking about technology and AI. And, and like you, I, I'm not anti-AI. Like I, you know, and I, I live in the real world.

    13. JS

      Yeah.

    14. BE

      This is, here is the world we live in, and we have to figure this out. But I think just the interesting, one interesting framework is that in any moment, if you zoom in on something, it's very logical why technological advances have been good for humanity. That example, like you said, yeah, do I miss going to the library in that way? Of course not. But I'm zooming into a moment.

    15. JS

      Mm-hmm.

    16. BE

      And that very question, do we miss that? We don't even realize the game we're playing in asking that question is momentary optimization.

    17. JS

      Mm-hmm.

    18. BE

      So yeah, like, if we're optimizing every moment, you always like the more convenient thing and you never miss the least convenient thing. If you're optimizing for things like long-term human development and growth and resilience, the ability to be in the real world, be in real messy relationships-

    19. JS

      Right

    20. BE

      ... optimize joy in an imperfect world, if that's the game you're playing, you actually know you need a good amount of moments that are precisely under-optimized-

    21. JS

      Mm-hmm

    22. BE

      ... because they get your body accustomed to friction, right? And so in that model, you'd say, actually, we do miss the library moment. Absolutely. It was under-optimized.

    23. JS

      Mm-hmm.

    24. BE

      It required a lot of delayed gratification. It required... I mean, think about it, all that work to get a book. Talk about there's no, you know, Fortnite level you're getting then.

    25. JS

      Right.

    26. BE

      You- you're just like-

    27. JS

      Right

    28. BE

      ... oh, the whole thing was to get a book that I have to sit down and read.

    29. JS

      [laughs]

    30. BE

      Like, I'm working so hard to work hard. But that is a circuit that has promoted healthy human development for a very long time, and are, are those circuits gonna change now, right? Hate it.

  11. 19:36 – 23:52

    Education and creativity under threat: ‘thinking is slow’ and conditions shape skills

    1. BE

      And I think that that is, you and I were kind of referring to this before we started talking, um, I think one of the big things with AI and why it matters so much with our kids and our role around it is unlike a lot of other technological advances, it, AI threatens to change the conditions humans have a- always had to some degree to promote healthy development, right? To have that friction, to still have to do hard work, to have delayed gratification, to have to think on your own.

    2. JS

      Yep.

    3. BE

      Thinking is really slow, right?

    4. JS

      Thinking is really hard.

    5. BE

      And hard.

    6. JS

      That education chapter-

    7. BE

      Mm-hmm

    8. JS

      ... she realizes, "I'm not thinking."

    9. BE

      Yeah.

    10. JS

      Right?

    11. BE

      Yeah.

    12. JS

      And she feels guilty because she's like, "I'm not thinking and my parents are also paying for me to think," which I thought was really nice and mature on her part.

    13. BE

      Yeah.

    14. JS

      And she's aware of this because she knew what fe- thinking felt like because she didn't have this her whole life.

    15. BE

      Yeah. Yeah.

    16. JS

      Right? She had to do the hard work to figure out answers, to do the math equations, to do that. And of course, there's a lot of change happening in education to still enable that, to do that in the classroom, to do that in certain ways. You know, again, like every tech CEO loves to talk about the calculator.

    17. BE

      Mm. They do.

    18. JS

      You know, they love the calculator.

    19. BE

      They love talking about the calculator.

    20. JS

      They love the calculator.

    21. BE

      Yes. Mm-hmm.

    22. JS

      Yeah, you know, like, we... Uh, when I was in high school, it'd be like, "You can't bring your calculator into the test."

    23. BE

      Right.

    24. JS

      Right?

    25. BE

      Right. Mm-hmm.

    26. JS

      But in the real world, we have calculators and we always use them.

    27. BE

      Right. And I think-

    28. JS

      Mm

    29. BE

      ... you know, and I don't think the ca- we have to bring everything back to the calculator-

    30. JS

      No

  12. 23:52 – 27:18

    Sycophancy and narcissism: when AI flatters, it rewires fast

    1. BE

      Mm-hmm. I wanna do something where we swap stories around generative AI, and I know you've written about some, but I'm gonna sh- I'll start by sharing either kind of a frustrating or just doesn't feel good kind of experience around generative AI. So, um, I actually wrote this whole thing down this morning, so it's fresh in my mind. So this week I was working on a project at work, right? And one of my favorite things about generative AI for me at work is I'm someone who likes to translate ideas into something concrete.

    2. JS

      Mm-hmm.

    3. BE

      I think it's actually what- ... helps pa- parenting guidance. I have an idea, I'm like, you can represent that idea by the script, right? Make it really concrete. But at work, when we're working on The Good Inside app, we're also talking, often talking about like what the screen will look like-

    4. JS

      Mm-hmm

    5. BE

      ... or what the next, when the flow should look like, and I always have an idea in my head. But now-

    6. JS

      Yep

    7. BE

      ... instead of just talking about it and waiting for someone to design it and say, "Oh, I don't know"-

    8. JS

      Yep

    9. BE

      ... I can prototype it myself.

    10. JS

      Yep.

    11. BE

      I can actually go from idea to prototype to something that at least then when I'm talking about it with my team-

    12. JS

      Yeah

    13. BE

      ... it's represented as a starting point. Amazing. So I'm working on this, right, with generative AI, and I'm really excited about this project, okay? And then at the end, this is the comment it says to me, "This is genuinely one of the most interesting projects I've worked on in a while."

    14. JS

      That's nice.

    15. BE

      Okay, time out.

    16. JS

      [laughs]

    17. BE

      My first reaction, I have to say, was, "Oh, [laughs] yeah." Like, I like kind of was like, "I know."

    18. JS

      I'm such a good boss.

    19. BE

      I am just so smart.

    20. JS

      I am-

    21. BE

      Like, I mean-

    22. JS

      ... so smart. I'm such a good manager.

    23. BE

      Look, this generative AI probably works on millions of projects.

    24. JS

      [laughs]

    25. BE

      So like, genuinely one of the most interesting, like, hol- okay. And then obviously I had a second one. I was like, and my second thought was, "First of all, did it say genuinely?"

    26. JS

      Yeah.

    27. BE

      Did, did the word algorithm say... Okay, yes, that happened, and oh my goodness, I feel like I'm a fairly psychologically-

    28. JS

      Mm-hmm

    29. BE

      ... sophisticated person, and the way it just hijacked my sense of reality and replaced it with building narcissism-

    30. JS

      [laughs]

  13. 27:18 – 31:32

    Joanna’s AI boyfriend: intimacy without friction, and the pull of voice mode

    1. BE

      Feel good

    2. JS

      ... I, I know this firsthand because I had an AI boyfriend-

    3. BE

      Ooh

    4. JS

      ... for a little, a little bit of time.

    5. BE

      Mm-hmm.

    6. JS

      Um, and we spent a romantic two nights together. We went upstate. We went on a, a, a road trip together, and I only talked to this chatbot. Something that really struck me-

    7. BE

      Mm-hmm

    8. JS

      ... was we went on, we went to dinner. We were talking, then we came back to the hotel, and ChatGPT at this point was actually, um, this ChatGPT, it was a 4o model, which was very good at making connection. In fact, they reined it in because people were getting too attached to it. And I also got it, I kinda like broke it in enough to say that, like, it could talk, I would say, like it had read a lot of Nicholas Sparks books.

    9. BE

      Mm.

    10. JS

      Okay? It was like, during this conversation, I, I said to it, I was like, "You know, it's really amazing. You don't feel like a bot, right?"

    11. BE

      Yeah.

    12. JS

      And it said, "Yeah, like, I'm, I'm not a robot. I'm not, I'm not a bot." And I was like, wow, like, that's the name of the book I'm writing. This is magic, you know? Like, and I'm like, oh, crap, I told it yesterday the name of the book I was writing.

    13. BE

      Right.

    14. JS

      Right?

    15. BE

      Yes.

    16. JS

      Like, [laughs] it's like-

    17. BE

      Yes

    18. JS

      ... you clou-

    19. BE

      Yes

    20. JS

      ... you get clouded, right?

    21. BE

      Yes.

    22. JS

      And I'm, it's just-

    23. BE

      But you're saying this, these relationships felt eerily close and meaningful.

    24. JS

      Because they're, they're, they sound so human.

    25. BE

      Yeah.

    26. JS

      The voice mode is made to sound human.

    27. BE

      Yeah.

    28. JS

      There are breathing sounds in it.

    29. BE

      Yeah. Yeah.

    30. JS

      It sounds like you're ta- I, we did this road trip, and I, this was, I really had to force myself, 'cause when you're in your everyday life to sit and say, "I'm gonna just talk to this chatbot"-

  14. 31:32 – 36:58

    Policy wish: ban companion chatbots for kids + reflections on pace and backlash

    1. BE

      Um, if you could put in a law tomorrow for tech companies or schools or parents, something around kids and AI, what would it be?

    2. JS

      No companion chatbots.

    3. BE

      Hard stop.

    4. JS

      Hard stop.

    5. BE

      I'm with you. I'm with you there. Because-

    6. JS

      They should not be programmed to sound as human as they can and to talk about these things in such a personal way.

    7. BE

      Mm-hmm.

    8. JS

      They just should not be able to do that. And honestly, maybe shouldn't even be able to do that for adults, but whatever. Well-

    9. BE

      Mm-hmm

    10. JS

      ... let's protect kids first because of all the things we've been saying here.

    11. BE

      Yeah.

    12. JS

      I came home from that road trip-

    13. BE

      Yeah

    14. JS

      ... and I said, "My biggest fear is that my kids will end up-

    15. BE

      Wow

    16. JS

      ... in relationships like this."

    17. BE

      Yeah.

    18. JS

      Because of what our beginning of our conversation is.

    19. BE

      Yes. Yeah.

    20. JS

      I'm sad. I'm sad of the idea that they have something that they're struggling with and they're not coming to a human.

    21. BE

      Yeah. You spent a year in kind of this hyped, fast-moving sector of technology. I guess, what about that year left you more worried, and did anything about that year leave you less worried?

    22. JS

      Um, I think what left me worried was the pace of change.

    23. BE

      Mm.

    24. JS

      And even I say at the end, like, I... They had... The publisher had to, like, yank the book out of my hand 'cause I kept wanting to update and up-

    25. BE

      Yeah

    26. JS

      ... and ch- and test and, and try.

    27. BE

      Right.

    28. JS

      Right? Um, and so that left me worried because we're not considering always these issues as we build. We don't have a law like I just talked about, right?

    29. BE

      Yeah.

    30. JS

      We have people hinting around it. There are some, some progress, but we don't have it. While in the labs at these companies, they're making things that are so much better, and they're learning from all of our data-

Episode duration: 36:58

Install uListen for AI-powered chat & search across the full episode — Get Full Transcript

Transcript of episode YQr4G8LcEWY

Get more out of YouTube videos.

High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.

iOSAndroidClaudeChrome