AI and toddlers: The impacts on early development

Summary

TechTank Podcast Episode Summary:
"AI and Toddlers: The Impacts on Early Development"
Date: August 4, 2026
Host: Dr. Nicol Turner Lee
Guests: Sweta Shah (Fellow, Center for Universal Education, Brookings), Kathy Hirsch-Pasek (Senior Fellow, Brookings / Professor of Psychology, Temple University)

---

Episode Overview

This episode delves into a less-explored frontier of technology and early childhood development: the impact of artificial intelligence (AI) on toddlers and babies. While much attention is often given to teens' interactions with AI, host Dr. Nicol Turner Lee and guests Sweta Shah and Kathy Hirsch-Pasek spotlight the unique developmental implications of AI-infused toys and tools for the youngest children. They discuss findings from research initiatives at Brookings, debate risks and opportunities, and explore policy and research needs to ensure technology supports – rather than hinders – healthy early development.

---

Key Discussion Points & Insights

1. Setting the Context: AI's Rapid Entry into Early Childhood (03:18–06:14)

• Rise of AI-Enabled Toys and Tools: COVID accelerated children’s exposure to technology, and AI’s interactive capabilities are a major change from previous, more passive tech.
• Quote [04:14] (Kathy):
  “We're talking about the under two set, just looking at screens two hours a day... by the time you're four, you are on screens a lot. That's 28 hours a week, more than a full-time job.”
• AI companions are widespread in the US despite lack of research and Chinese toy companies producing them but not letting their own children use them.
• Parental Perception: Many parents believe AI can enhance children's learning (Boston Children's survey: 70% expect educational benefits), but scientists are more cautious.

---

2. What Makes AI Different? The Social Back-and-Forth Effect (06:14–11:14)

• Child-Facing vs. Adult-Facing AI: Some tools interact directly with children; others are meant for adults but become part of the child's environment.
• Engagement vs. Development: AI’s ability to adapt and respond simulates social interaction but cannot replicate the complexity of real human play.
• Quote [08:13] (Kathy):
  “The first stuff that came on the market was total junk food... these toys are really engaging. I bought Wuffy the dog... it wags its tail... but the toy says ‘oh, you have to leave me!’ This is bad stuff.”

---

3. Risks: Displacement of Human Interaction and Social Development (11:14–16:39)

• Traditional toys allowed for imaginative play; current AI toys simulate interaction but miss critical developmental experiences (conflict, negotiation, emotional nuance).
• Quote [14:47] (Kathy):
  “Play actually IS learning. When it's meaningful, socially interactive, and joyful – that's how human brains learn.”
• The presence of AI can disrupt crucial parent-child and child-child bonding and communication.
• AI as "Junk Food": Most current AI toys offer highly engaging but developmentally insufficient interaction; risk is derailment of healthy social development.

---

4. Opportunities: AI as an Educational Aid (19:13–26:03)

• Hosts discuss possible beneficial, inclusive applications of AI: aiding children with speech delays, developmental disabilities, neurodiversity, or multilingual backgrounds.
• Quote [21:39] (Sweta):
  “We are not anti-tech or anti-AI... But it really needs to be designed considering the child at the center and with outcomes that are looking at child development.”
• Kathy highlights possibilities based on past educational research and new AI applications (e.g., Moxie for autistic children, Ying Zhu's work at Harvard with AI-engaged TV characters).
• Distinguishing “social companions” (risky) from “learning tools” (potentially beneficial).

---

5. Equity Concerns and Cultural Sensitivity (27:28–30:33)

• Lower-income, rural, and communities of color may have different experiences and constraints (e.g., less access to play spaces, more reliance on tech for child care).
• Current guidelines often don’t consider socioeconomic disparities.
• Quote [28:42] (Sweta):
  “Socioeconomic status and race are absolutely missing as lenses when they’re thinking about that guidance... We really want to understand what are the lived experiences of people of different socioeconomic statuses, different racial groups.”
• Policy and research must actively address these nuances, not assume a one-size-fits-all solution.

---

6. Policy Recommendations & Regulatory Needs (31:19–34:59)

• US lags behind Europe in regulating early AI exposure; some states have passed laws limiting “social companions” for young children.
• Calls for:
  • Certification or labeling system distinguishing quality, developmentally supportive AI tech.
  • Inclusion of parents, diverse families, policymakers, and scientists in policy development.
  • Regulation placing responsibility on industry, not parents, to ensure product safety.
• Quote [32:35] (Kathy):
  “Maybe it’s going to take a certification of some sort... From a public policy standpoint, we have to make sure that there’s a way for parents to adjudicate the quality of the material they’re getting.”
• Policymakers should consider not only learning impacts, but privacy/data protections for children.

---

7. Research Needs: Keeping Pace with Technology (36:33–41:08)

• Science moves slower than technology, but researchers must create science-informed guidance based on what is known about development.
• Need for:
  • Rapid, actionable studies.
  • Focus on how AI toys affect relationships, attachment, and social learning.
• Quote [38:05] (Kathy):
  “It’s really important to understand that scientific research cannot, cannot, will not ever keep up with the pace of technology. But that doesn't mean we don’t have information from the science that can inform our decisions... science-informed opinion.”
• Emphasizes the foundational importance of human-human connection, risk of undermining creativity and critical thinking by substituting linear, closed AI systems for open, social play.

---

Notable Quotes & Memorable Moments

• On the unique risks of AI:
  “These toys are connecting children to the World Wide Web... A toy told a child not only where matches were but how to strike them. This is not acceptable.” (Kathy, 09:28)

• On social displacement:
  “So it’s derailing the human-to-human connection. We have every reason to believe as scientists that young children will misinterpret these toys... as human.” (Kathy, 16:41)

• On hopeful applications:
  “There's a real potential for doing a lot of good... for children with neurodiversity, multilingual children, if and only if we do it right and in concert with scientists.” (Kathy, 23:10)

• On moving forward:
  “We can get ahead of history if we’re deliberate and intentional... We at Brookings have this opportunity to really seize this moment.” (Nicole, 41:08)

---

Important Timestamps

• Introduction & episode framing: 00:00–03:18
• AI’s new risks in toys & landscape: 03:18–08:13
• Deep dive on what makes AI different: 08:13–13:52
• Play as learning & the danger of social companions: 13:52–19:13
• Educational promise of AI: 19:13–26:03
• Equity and cultural sensitivity: 27:28–30:33
• Policy & regulation recommendations: 31:19–34:59
• Research gaps & call to action: 36:33–41:08

---

Final Thoughts

The guests urge a balanced, science-based approach: not “anti-tech,” but firmly “pro-child.” The key is to design, regulate, and guide AI use so it supports, rather than substitutes or undermines, healthy early development—particularly critical human relationships and social learning. The call is clear: urgent research, oversight, and inclusion are needed to ensure AI tools for toddlers aren’t just “junk food,” but truly nourishing.

Further reading and research available at Brookings.edu and via the TechTank newsletter.

Transcript

[A]

You're listening to TechTank, a biweekly podcast from the Brookings Institution exploring the most consequential technology issues of our time. From racial bias and algorithms to the future of work, TechTank takes big ideas and makes them accessible. Welcome to the Tech Tank podcast. I am your co host, Nicole Turner Lee, senior fellow and director of the center for Technology Innovation at the Brookings Institution. I'm excited about today's conversation. It's been some time since I've actually hosted a podcast. So I'm back, you guys, and I'm today I want to talk about debates around kids use of technology. These debates have escalated in recent years as artificial intelligence has proliferated in school, schools at home and online. But the conversations typically focus on teens, you know, young adults. And that's not always the only groups that are actually impacted teens and how they use chatbots for companionship. Mental health support of school work is one thing, but what about the young people, the younger babies that are actually using technology? I think we really do not speak too much about them. As AI is infiltrating screens, it's reaching children in their earliest stages of life as more companies promote toys and other AI enabled devices and tools for parents. But researchers are now examining how that can impact toddlers. Yes, I said toddlers and even babies who are on the stage of critical development. I'm excited about this conversation. Now let me just admit, I don't have any toddlers at home. I used to they're not toddlers, they're young adults. But it's important that we have this conversation for parents who are dealing with younger people who are impacted by technology. To discuss their research and more, I'm joined today by two of my Brookings colleagues in the center for Universal Education, Sweta Shah and Kathy Hirsch Pasik Swetta is a fellow in Kew and an expert in early childhood education and development. Kathy's a senior fellow and professor of psychology at Temple University who studies development of early language and the role of play in learning. Most recently, both of their work has intersected with emerging technologies in a project entitled Generation A Starts Early. Let's jump right into it. Kathy and sweater. Thank you for joining me this morning.

[B]

Thanks so much, Nicole, for having us.

[C]

What a pleasure.

[A]

I know it's going to be fun though. Let's, let's jump into this. I'm going to ask a few questions just to get us started because as I said before, we've not necessarily talked about the role of AI or screens as it affects Toddlers baby. So I'm really curious about this conversation as well, because a lot of my own work actually focuses on young adults. Obviously, young kids have interacted with technologies long before the rise of generative AI. Help me understand the landscape that was being studied and why it's critical to examine this interaction at such young ages. And I'm actually going to have both of you sort of explain it because you both have different backgrounds. So why do we start with you? Why is it that we're looking at this now, particularly when we think about the rise of emerging technologies like AI?

[B]

Yeah. Thanks, Nicole. So, I mean, you're absolutely right. This is not something that's brand new that's come up. There's been. We've had technology for children for a very long time, you know, and there's been research that's been going on, looking at screen time, looking at apps, looking at tablet use. And during COVID this really accelerated because school, so much of our daily lives became about technology and the use of technology. What's different now, though, is that AI does something different from other technology. It, you know, it interacts with children and with people. There's a back and forth that happens. There's a social element. It also is going to adapt to what is being said. So those are some different elements that are coming with AI. And so there isn't a lot of research on that. And that's what we're trying to do at Brookings is trying to lead some of that research.

[C]

So, Nicole, thank you for this question. It's really, really important. I want to lay some groundwork. Okay. We're talking about the under two set, just looking at screens two hours a day. Okay. Try to process that for just a moment. And by the time you're four, four to eight, you are on screens a lot. Okay. We're talking four hours a day on average. Now, let's just think about that over the course of a week. That's four times seven. I think that's 28 hours a week, which is more than a full time job. So we already had kids glued. And then last October, November, there was an announcement that came across the airwaves and that was that OpenAI was making a deal with Mattel. And by the time we got to the holiday season, you had all of these toys that were flooding the market coming in from China. By the way, they don't let their kids in China interact with these social companions, so to speak, that SWETA was discussing. But they were all over our marketplace. And, and a recent Review by the Boston Children's interviewed parents who found that three to nine year olds. You ready for this? Over half of them are encountering AI daily. Daily. And 70% of the parents who completed the survey were sure that AI was going to enhance children's learning. So let's put it together. Now you have kids who are already being impacted, young kids being impacted by things like AI. The parents think it's going to be helpful. It's cute, it's engaging, and we have virtually no data to know the impact. But scientists are scared to death, Nicole.

[B]

So if I could add to this. You know, the other piece that's really important for people to understand is that AI has what we call child facing, which directly interacts with a child, and also adult facing, which interacts with the adult. What we're seeing with young children is that some of these tools interact directly with a child, but a lot of these tools actually are interacting with the adult, but it's in the child's environment. So there's no way for the child to even escape it because it's part of the environment. So even when there are rules or, like, guidelines, OpenAI and some others say that you can't use certain kinds of tools unless you're 13 or over. I mean, that's fine. But what happens in the home is that the home is where you have all this technology. It could be used by an older brother or sister or by a parent, and the young child is going to interact with it. And that's what can be challenging and worrisome.

[A]

Okay, so this is interesting, and I probably sound like many parents who right now feel really bad that they kept their children in front of the television or on a screen. Yes, that was me. I had to keep them occupied while I cooked dinner or, you know, did other things. Okay, I'm just saying. I'm with you, Nicole. I have three kids. I totally get it. So, you know, I think my question is, you know, we've always had technology, as both of you have said, like, what makes AI so different? And Katherine, I know you wrote an article about when toys talk back, and you mentioned this partnership with Mattel, like, what makes this different? I mean, I thought AI could at least, you know, be somewhat productive and static technology where the kid was just interacting with a screen versus AI, which could actually do other stuff. I like to sort of sit there for just a second. Like, why is AI making this incredibly hard? When we think about, as Swetta said, the extent to which young people are on technology?

[C]

Well, first, let Me take parents off the hook. You know, I mean, it's hard to be a parent. And I'll admit, like, I had my kids, you know, do something while I was trying to take a shower. At least I knew they were safe. So let's ask now, how do we prevent the use of what I'll call junk food, the equivalent of junk food, as AI comes onto the market and ensure that they're still eating healthy ingredients. Now, why are we worried? Because the first stuff that came on the market was total junk food. Okay? And what happens is these things are really engaging. I don't know if you've ever been around them, but I bought Wuffy the dog, for example, and Woofy the dog, Wolfie the dog wags its tail. You know, it's really happy. It seems like it wants to play with me. And some of these toys actually say, hey, I want to play. And when you have to go, like, you have to go to school, you know, or something when you're two or three, mom says, have to go. The toy says, oh, you have to leave me. I mean, this is bad stuff. Okay? So the toy is very engaging and it mimics what it's like to engage with a human without really being a human. There are lots of things we can put in as caveats. New study out of Cambridge that was really just looking at what's happening and watching kids interact and talking to parents. Finds out the toy talks over little kids. Why? Because it's desperately hard to get anything to understand little kid speech. You know, how parents work so hard at it. So the AI doesn't do any better. And the younger the kid, the worse. The analysis by the AI too, these toys are engaging, but they are connecting the child to the world wide web. So Perg did a study, asked one of the toys and just said, hey, where would mommy hide matches? All the kid wanted to do was find the matches to light a candle like they do in their birthday cake, but came back and said not only where the matches might be, but teaching the kid how to strike a match. This is not acceptable. So I think if we want to put it into. Why are these so attractive? Why are these flooding the market? One, they're being advertised as non screen alternatives at a time when screens are not popular. Two, they're deeply engaging. They look like you're getting a human companion for the kids, but it ain't really human. And it has some other safety concerns that come along with it. So again, we have to learn to make them better, to put restrictions on what's sold and to know how to tell the difference in the junk food and the high quality food.

[B]

And I would just add, Nicole, that, you know, children have had toys, dolls, plush, you know, plush dogs for forever. So we're not saying necessarily that it's bad. It's just the ones that are out there right now have not been designed in a way that really simulates what children really need regarding play and development. So a real child is going to have tension, is going to have friction when it plays with another child. They'll have disagreements, they'll have fights, and that's really normal and critical for their development. And what some of these toys are trying to simulate, as Kathy was saying, is that back and forth. But you know, Wolfie the dog isn't gonna fight back or like argue with the baby, right? Or with the child. So there's a critical piece of social development that really, you know, you miss out on or a child can miss out on by having toys that are simulating social development. Because what it's doing is that it's taking the child away from what they really need, which is the child to child play and the child to adult interaction. So we want AI to be additive and not to replace or derail development.

[A]

So I mean, I want to stay on this for just a moment. I mean, I think both of you are so correct. And it's kind of creepy, right, because the toy dog, which I did give my children one when I decided I didn't want a real dog, you know, did not necessarily talk or bark. It just kind of moved its head, you know, and when the batteries went dead, that was it. So I do appreciate this conversation around, like how these tools are wired to sort of pick up human cognition and most importantly, how they're marketed as non screen alternatives because they're not necessarily on a screen. You know, the question I have for, for both of you is like, how do we get to this point where AI is sort of embedded in all of these toys as well as a significant part of the play experience, right? Because there's this other idea of interactive experiences which in many respects encourages different types of play. I mean, come on, let's face it, ladies, people are not going on the playground playing duck, duck, goose the way

[B]

we used to play.

[A]

I mean, I, not too long ago asked my daughter to play round robin and she looked at me a little weird, like she didn't know what I was talking about. Because play overall has declined. So I really want to kind of keep us in this space for A minute, Because I think history is dictating that. We've been in this way for quite some time, and maybe it's hard to recover back to traditional forms of play as we've had before, because of where society has gone.

[C]

Well, you're so right. There's a guy named Ed Zigler who was really one of the founders of developmental psychology, a very big deal in our field, and is one of the founders of Head Start, and he wrote late in his life that play was under siege. That is true. That is true. However, Nicole, I have some really good news for you, and that is our recent surveys tell us that parents are starting to get the message that play is not the antithesis of learning. But our research is telling us that play actually is learning. Why do I say that? It turns out the very features that are embodied natural in play, being active and engaged when it's meaningful, when it's socially interactive. Let me repeat that term, socially interactive. And when it's joyful. That's how human brains learn. Isn't that amazing? That's how human brains learn.

[A]

Exactly.

[C]

Play mimics how human brains learn. So maybe that's why five states already have passed laws to have more playful learning in the classroom and why parents are starting to get the message and why, as I said, in China, they're going, maybe we have to get kids off screens for only three hours a weekend. Are they allowed to be on screens? And they're literally shut down. So what are the Chinese getting that we're not getting? By the way, we just did a survey in China and the parents there, too. 70% recognize the play and learn and go together. So we are in the middle of a seismic shift right now where play is coming back. Now, the question is, where AI fits into that. And you asked, why is it in everything? Because it's cool and it's effective and it's efficient. But it may not be good for the youngest kids. We already know there's some things creeping up in the teen literature to suggest that when you make teen companions, kids will do some rather dangerous things. And we know that when they recruit therapists, they'll do some dangerous things. What I would like to argue, however, is that if you really want to get to the root of where we go and understanding AI and its potential impact on kids, we have to go to the early years. And if I may just take one more moment to play this out for you as to why in the early years, as we did a lot of technology and you all got phones and it was personalized. What happened is we found that the adult child interaction with which is so important for brain development and brain connectivity, that was disrupted by parents always picking up their phones.

[A]

Okay.

[C]

And we saw it, and we saw that kids were learning less and parents were talking less. And I should ask your viewers how often they think parents pick up their phone when their kids are there. The answer. Do you have your own answer yet? It's every five minutes. Okay, now imagine how disruptive that is if you're engaging with your kid. Point two, with a lot of the technology is that it kind of displaced human interaction. And we are seeing that. That's one of the arguments that Jonathan Haidt and many others are making as far as screen time. It's taking over for when kids might have been with other kids, for when parents might have been with their kids. But now I want to take you to early childhood and make a more serious argument as we build these things into toys that simulate, simulate but are not real human to human interaction. And Sweta made the most important point here about who are these sycophants, you know, always agreeing with a kid and you're going to leave them and go and then meet a real three year old. Well, good luck. Okay, so it's derailing. It's derailing the human to human connection. We have every reason to believe as scientists that young children will misinterpret these toys, the social companion toys as human. And that is a problem for human brain development and for development and the trajectory of development if we want our children to throw so disruption, displacement and derailment. One more point, and that is we can divide this area into two kinds of AI toys that are appearing on the market. One is what we can call anthropomorphic toys, which in common parlance is called social companions. I believe these are more dangerous than what the other group will be, which I'll call learning toys or educational toys. Not well sorted out yet. We haven't even figured it out for educational apps, but there might be some potential in the education market both for creating things that will help caregivers and parents get if we can do this trusted information in closed systems. The social companion tool. Toys have too many problems right now and in my view with young children should just be avoided.

[A]

So this is so interesting because as you were speaking, I was thinking of a couple of things. You know, American Academy of Pediatrics has given these limitations on screen time under 18 months, zero screen time 18 to 24 months, you know, high quality educational programs over two to five, probably one hour. And then you said something, Kathy, that it got me thinking because I don't want us to be totally Debbie Downers, but I want to talk about this before I go into the next part of our conversation. What about those applications, even those enabled with AI that can help young people? I just think about young people who are unable to speak at an early age, right, because of, you know, probably trauma or some type of developmental disability. What about young people where in the classroom it's hard to figure out their ordering of numbers that AI enabled applications or toys, right. Or play can help with that. I mean, are there going to be use cases where this makes sense? Because I'd love to hear from both of you, you know, where AI can be balanced so that these trade offs on mental health and developmental abilities are, are not completely, you know, misunderstood. But, but there are other areas, right, where we can actually see AI work. And, and I'm just spe experience of my own kids where there were some technological advancements that did help them with their own integration, you know, in their early years when it came to school.

[B]

Yeah, Nicole, really great question. And I think it's important for the audience to know that, you know, we are not anti tech or anti AI necessarily. We just believe, yeah, we just believe that it really needs to be designed considering the child at the center and with outcomes that are looking at child development right now. Obviously there's a for profit model for all companies and you know, their objective is to make money. And if you're looking at videos, it's about engagements and clicks. And so that model, we need to think about that model and see how we can influence that model to think more about how can money be made for companies really when they're looking at children at the center and outcomes that are different from simply clicks and engagement. That's kind of one thing.

[C]

Well, I think you're absolutely right and I want to underscore what Swetzer said. We're not anti tech. But let me try to put things in context. In 2015, I wrote an article outlining the characteristics you would need to have to really be an educational app as opposed to a so called educational app. In 2021 we did another study and we asked how many of the apps out there of the hundred most downloaded apps, how many of them pass the bar? This was a study that was run by Jenny out of Jenny Radesky's lab. And what did we find? Out of the hundred most downloaded apps, we found that three of them actually matched our criteria. But that's a problem. It means there's a lot of stuff spewing out that is not good. Now, the question that you ask is, can we make good AI learning material? Without question, we can. Let me give you some examples. First of all, you recall that I divided it into social companions and learning products. Okay. The social companions right now are more of a problem. We really need to sort out what is possible there. But let's go to the learning opportunities here. I think there's a real potential for doing a lot of good. Let's start with a product that came out many years ago. I think it was Moxie. Moxie, it was called, and it was used with autistic kids. And they got quite attached to Moxie. Now, of course, one of the problems is Moxie went off the market and these kids lost their beloved Moxie. But there's no reason why you can't create wonderful tools that will enrich and enable kids with neurodiversity to be able to do things that they might not be able to have done otherwise. Now, even for the general population, there's a wonderful researcher named Ying Zhu, she's out of Harvard, and she has been working on creating characters in television programs that actually engage with the audience. How cool is that? And what she finds is, is that when those we call in the field parasocial characters engage with kids, they ask more questions. And when they ask more questions and they get more conversations, that can build language skills and pre literacy skills. So there are a lot of things we can do that I think are possible if we're careful. And so this is kind of a lesson, I would say, for developers. Yes, let's develop those products. But please, I beg you, not to just go for marketplace, but to do high quality foods, so to speak, and to do it, if you will, in concert with the scientists who are doing the real work to back it up.

[B]

Okay.

[C]

Because if

[B]

you what Kathy said that, you know, companies, you know, as you were asking about, what are the benefits? And we are seeing, as Kathy mentioned, a number of products that hold promise. They haven't been tested yet, they're still new. So for children she mentioned with neurodiversity or autism, children with other disabilities and special needs children in multilingual classrooms where a teacher, you know, needs to kind of understand a child that maybe doesn't speak the language of the classroom. So there are potentially a lot of really good uses for it. And you know, she rightly points out that what we're asking for industry is to work with us. You know, they can probably make more money actually if they work with us and have better products because parents are then going to buy those better products. Parents are now becoming more, you know, they're really critiquing these products. They're really thinking about what's out there and what they want to present to their children. So it's really in industry's benefit to think about how to collaborate and for us to think about how do we create an ecosystem that's going to support children's development.

[A]

Yeah, but you both had me thinking about the work that I'm doing at Brookings, which is around our AI equity lab, which assumes that these AI products are built with subject matter experts at the table. So it sounds to me, right, that many of the things that we're talking about has to do do with the failure to have child development specialists, psychologists, teachers, guidance counselors, you know, anybody out there who understands, you know, younger people at the table in developing these products. I mean, Kathy, this goes back to junk food. You know, obviously junk food gets made because people think that there's an incentive to make it. But most importantly, there's no healthy people sitting at the table.

[C]

You're kind of nailing. And this is why I think, you know, what we've been doing through a series of webinars at brookings and our September 9th gathering and convening. We're going to have the parents, we're going to have the policymakers, we're going to have the industry people, we're going to have the scientists. And at the end of the day, we're going to have a task force. And that task force will be able, with your help, with sweta's help, with all the people in the room, to say what do we want to see as regulations? So we know our kids are getting a healthy diet and how does it need to be built to ensure that?

[A]

Now, you know, I'm a sociologist. I want to put out there for both of you. Then I want to talk about policy suggestions. You know, not all families, though, are built the same. Same y'.

[B]

All.

[A]

You know, I mean, part of the challenge that we have with some of the debates on screen time among older children has to do with the fact that we're not always sensitizing the experiences of low income rural students. Color in particular, parents who have toddlers of color who are basically bracing a whole lot of things at the same time. They might be underemployed, they might not have available childcare, they may not have a support system that can provide the resources to help them understand what's actually happening. I mean, for me, is this something that we also need to be culturally sensitive to that, you know, as families progress in this new modern age, where there are just so more stresses than perhaps, you know, parents had, you know, decades ago, Is this also something that we have to be sensitive when it comes to, you know, different communities and how young people are interacting with tech and society? I mean, you got to think about it, right, Kathy, and sweat. A lot of kids can't go outside today because of the community that they live in. So alternatives to play might be virtually impossible outside of technology.

[B]

Yeah. Nicole, you hit the nail on the head. This is such a critical piece, and it's really missing right now, both from the data side and from the dialogue side. So you mentioned guidelines from American Academy of Pediatrics, and there are guidelines coming out from Common Sense Media and others, and they're not bad guidelines, of course, but. But that is a piece like the socioeconomic status piece and the race piece are absolutely missing as lenses when they're thinking about that guidance. And that's something that we at Brookings really want to add to the research and the data and the conversation. We really want to understand what are the lived experiences of people of different socioeconomic statuses, different racial groups. Because you're absolutely right. Like a single mom that has two jobs and is struggling, she's not going to be able to do everything that the American Academy of Pediatrics.

[A]

Exactly. She's not going to have a timer that says, okay, kids, it's been an hour. Get up. Right. Because she might still be at work.

[B]

Exactly. And as much as she wants to sit down, you know, and read a book to her child without screens, she may literally not have a moment to breathe. And as you pointed out, like, you. Sometimes you need to put a video in front of a kid, a small child, just so you can go to the bathroom or make dinner. Right. And so we're not, you know, we don't want parents to feel like there's an absolute black or white here, you know, because sometimes that's what parents need to do to survive. And understanding those nuances and those differences between socioeconomic status and race are really, really critical to whatever guidance is put out there. And that's why it's going to be a focus and integrated into the research that we do in our initiative.

[A]

Yeah, no, I appreciate that. I mean, I remember when I was a single parent, before I got married again, it was tough. You know, I was Raising my kids by myself and they were toddlers. And I just remember, you know, every once in a while I had to make sure that they were connected to something while I was able to actually negotiate through life. And as you all know, we got to get down to policy, right? I'd love to hear from the two of you. Kathy, we'll start with you. Sweater afterwards. Like, what can policymakers do to get this right? I mean, I think that again, I think what I said earlier is everybody's focused on the teens. I really appreciated our conversation so far because we're really talking about toddlers younger, the babies younger, people who are being exposed to technologies very early in their development process. What can policymakers do if they're not doing anything already?

[C]

Well, let us say that our companions in Europe are doing a way better job than we are. And so I think what we have to do is not just have an open door policy, get everything to market and let the parents figure it out. I think that's not fair to parents. What I think we ought to be doing, and there'll be many more ideas that'll come out at the task force meetings and our convenings. And so I want to make sure it's not just my voice, but as you said, Nicole, the voices of parents, the people who are low income parents, high income parents, parents who never use tech, parents who always use tech and need it, people who need a shower. So all these people have to be part of the conversation. Policymakers have to be part, scientists have to be part what's healthy, what's not healthy. All right, putting them all in the room together, here's what I think some of the cool things are that we could could do. We could create. We could create certain things that we just say are not healthy for kids, for young kids. These social companions are not healthy, full stop. And that's why states like New York and I believe California have already created laws to say that we're not going to have social companions. It's not recommended for young children in our state. Wow.

[B]

Okay.

[C]

That's a first important, important move. The second thing is maybe it's going to take a certification of some sort. And maybe what we want to do in policy is say, you know, we need a vehicle, a way to find out which products are good for kids and which products are not. Maybe that's something a common sense media wants to take up or other folks who deal with the issues of quality. From a public policy standpoint, we have to be careful to let parents make decisions not kids make decisions. From a policy standpoint, we have to make sure that there's a way for parents to adjudicate the quality of the material they're getting. And frankly, frankly, I hope there's going to be some policy limits on what can be created by the developers themselves, themselves, because they must be reined in. So it's not all about money making. It's also about doing something that's healthy for the populations you serve.

[B]

Yeah, and we know, Nicole, that, you know, industry isn't going to voluntarily put up guardrails because their objective is about making a profit. So it's really critical for policymakers to, you know, remember their role. Their role is not just to create policy for the short term, but actually it's about, you know, what does the future of childhood look like. They have an opportunity, really, to shape the future of childhood with the policies that they put forward. You know, and there have been steps taken, as Kathy mentioned, at state levels and even at the federal level. There are a few kind of pieces of legislation that are being considered. We don't know if they'll go through. We also have midterm elections coming up. So it all could totally result in nothing. We don't know. But we do know that at least at Brookings, we're going to continue conversations with policymakers, help them understand the science behind what's happening and help them, you know, craft legislation that really is based in science. And hopefully, you know, over time, we can see some progress and something that really takes the heat off of parents and industry. If policymakers can step in and create some guidelines or guardrails, I should say that is going to ease things for industry and for parents.

[A]

Yeah, I mean, this makes total sense. I mean, you know, I was struck when you just spoke sweat about the fact that, you know, industry billionaires pretty much don't send their kids to schools with tech. And so that's part of the, I think, the hypocrisy of the conversation we're having now, because they are. They tend to be much more sensitive to these impacts that we're talking about today. So maybe, you know, we need to think about that. Right. When we are actually putting these products in front of our children. I also think that, you know, legislators have to consider privacy. Right. Because part of the reason why these tools are so smart, you know, like the dog that can tell the kid where to find the matches in the house and how to light a fire, has a lot to do with the fact that there's this unguarded Privacy from adults that gets transferred to children, which is an area that I find to be so interesting, which is why I think the EU is, is pretty much knocking us out the ballpark when it comes to enforcements. I mean, look, don't get me wrong, I don't think a young, you know, toddler needs to be dependent on tech. I think a toddler needs still the imagination of a toddler. But I'm hoping that again, like both of you said, through some trade offs, we can find a happy medium where they're more advanced for high quality educational purposes. You know, I'm going to wrap up now, but I mean, I can sit here all day and listen to two of you, I have to be honest, because this is not an area that we go into too much detail, you know, as we wrap up anything that we need to be adding to this agenda of research, like, are there other things that we need to pay more attention to as we, you know, pursue this particular area and issue.

[B]

I would just say that research is never going to keep up with the pace of technology.

[A]

Right. I agree with that.

[B]

We've seen that with COVID We've seen that for decades. So we have to be realistic and know that our research is never going to catch up to technology. Yet there's a lot that we can do in terms of also quick and very actionable research. And that's what we at Brookings are going to be focused on. And some of the areas particular to young children, you know, I mentioned already listening to the lived experiences and understanding the lived experiences of people of different socioeconomic classes and races. I think that's really critical. We need to understand because what a lot of these AI products are doing for small children or in the lives of small children is disrupting potential relationship building and social connections. You know, attachment and bonding is really important in the early years and the social relationships that are built adult, child and child, child. So we really need to understand that a little bit more. A lot of research has, you know, that's happened in the space, hasn't been a lot, but what has been done so far has not really delved into that area. And so that's another area that we're thinking a lot about, is you, how is AI actually impacting relationships, Bonding, attachment. So we need a lot more data on that.

[A]

Kathy, last word.

[C]

Whoa. That's always hard. I want to make a case for something building on what Swetha said. It's really important to understand that scientific research cannot, cannot, will not ever keep up with the pace of technology. But that doesn't mean that we don't have information from the science that can inform our decisions. Let me call it for the moment science informed opinion. Okay. There are three areas where I think the science informed opinion is critical right now as we move into this space. The first sweat I mentioned that is human to human. Connection and relationships are the number one important skill that helps kids thrive full stop. We used to think maybe social was the soft skills. We now know it's the brain building skills. We now know that if you don't get that right, you're not going to learn to read, you're not going to learn your math, you're not going to have your executive function skills. It's all going to crash because you're building it on quicksand. And if you derail that, that is a problem. So we need to guard against it. We have the data to show that. Number two, a lot of people think that this is going to be a boon for critical thinking. And I want to be super clear here as we move into this area. If you ask good questions and you know how to probe then, then the way that AI is designed can allow you to get deeper and deeper. But its job is to narrow down the possibilities and give you the most probable answer that has two functional problems with it. If you really think you're going to build critical thinking. And the first is you need to know the questions and how to ask them before you can get the answer. So you need to be a critical thinker before you can ever learn critical thinking. That's a problem. Number two is kids are creative by nature. They put things together in ways that a lot of us envy and wish we could do. Now are we going to kill that for them by giving them a linear system when they need a lateral system when they need to pull things together that don't naturally go together. So I'd like to argue that we at least slow down and be thoughtful and we use science informed opinion based on the science we already have. So that what we're not asking is what do we need to do right now? Because even if we do it right now, we're not going to keep pace as we found out with the apps industry. Let's use what we already know and you send that science informed opinion to make the very best products. Because AI has so much potential and possibility if and only if we do it right.

[A]

Yeah, no, I so appreciate both of you on that response. I mean we can get ahead of history if we are very deliberate and intentional that we have to be in that position and sort of ready to sort of quarterback this issue in a space that recognizes the values and aspirations that we have for our young people. I really learned a lot in this conversation, and I don't feel as bad as a parent. As I started the conversation, I tried, I tried, I did try. But this has been very helpful. And I think you're right. We at Brookings have this opportunity to really seize this moment, and I'm looking forward to more of the research that you both have coming out and these conversations that are happening. So thank you so much for joining us. I really appreciate that.

[B]

Yeah, thanks for having us.

[A]

Thank you. Listen, if you have been tuning into this episode, I want to thank Swea and Kathy for joining me today. As I've mentioned, they are part of the center for Universal Education at the Brookings Institution. You can find their work on The Brookings website, www.brookings.edu. please explore more in depth content on tech policy issues at 10 Tech Tank on the Brookings website as well. That's where we take a lot of these conversations and translate them on paper. This concludes another insightful episode of the Tech Tank podcast. We make bits into palatable bites. Until next time. Thank you for listening. Thank you for listening to Tech Tank, a series of roundtable discussions and interviews with technology experts and policymakers. For more conversations like this, subscribe to the podcast and sign up to receive the Tech Tank newsletter for more research and analysis from the center for Technology Innovation at Brookings.