Maggie McKay (Host): Welcome to the Parent-ish Podcast, where experts at Children's Mercy, Kansas City talk about the little everyday things parents experience with their babies, teens, and in-betweens. As a parent, how much do you know about chatbot interactions with children?

Mark Connolly, Professor of Pediatrics, University of Missouri, Kansas City School of Medicine, joins us to share his expertise on this very important topic.

Thank you for being here.

Mark Connelly, Professor of Pediatrics: Thank you for having me.

Host: Can you explain what chatbots are and how children are encountering them in their daily lives?

Mark Connelly, Professor of Pediatrics: Sure. Actually quite timely issue. There was just a recent survey done by Common Sense Media that suggested up to 72% of adolescents are using AI chatbots for companionship now, with half of those doing it daily.

Host: Wait for companionship?

Mark Connelly, Professor of Pediatrics: For companionship specifically. So it was kind of an interesting to ask about that and then we can kind of talk about that as well. But yeah, I think the main thing to recognize with, with chatbots here in particular in the last few years is they're based on a form of artificial intelligence, that gets trained on language. So it basically gets trained on almost everything on the internet and is able to kind of predict, what do they think you're asking for in a prompt. And so they're trained on really kind of fancy mathematical models and they get really good at this point of being able to understand patterns of speech and pattern and then speak back to what they think you want, based on whatever prompt you give. And then again, to that point, to the companionship. So prompts around that I'm feeling lonely, the chatbots are able to kind of recognize what the sentiment is of that, use their information that they've been trained on to come up with some sort of response that they feel like is appropriate for that particular context of language.

Host: Oh my goodness. Okay, Mark, I'm already scared, as a parent. Are there concerns about children forming emotional attachments to chat bots?

Mark Connelly, Professor of Pediatrics: The challenge with that question is there's just not a lot of science quite yet because these models are so new and particularly again, the kind of what's called generative AI chat bots. So the much more creative, much more fancy chat bots that they can kind of really recognize large patterns, and speak back in really kind of random ways. So we don't really know what these kind of newer models have as an impact on kinda social development and emotional development. But, the thought is similar to other things with adolescents and our younger kiddos, is they can form companionship pretty quickly with things. Even if you think about your or your younger ones, sometimes they, they think of, human-like characteristics in just kind of objects around them. So if you can kind of apply that then to digital tools and digital companions as well, the idea of kind of forming true, kind of lasting emotional bonds with these type of chatbots.

I think there's concerns that that does happen and potentially replaces what we still view as kinda healthy relationships that develop with humans ultimately.

Host: Yeah, this is next level imaginary friends for sure. How do children perceive the quote unquote intelligence or authority of a chatbot? How do kids see that?

Mark Connelly, Professor of Pediatrics: One of the things for parents to know about chatbots as well, and one of the major limitations of what happens with chatbots is because they're fairly creative on how they generate response, sometimes they do generate responses that sound really confident, what's actually called a hallucination interestingly.

So they generate a response that's actually factually incorrect, but sounds really confident. And of course if you pair that then with kinda what we know about adolescent development, they can really attribute that as kind of an authoritative type of response and assume that that's correct.

So there is some evidence that adolescents really do view this information as accurate that comes outta chatbots and younger kids as well. And so there's obviously concern that if sometimes they're generating potentially problematic or false information, or even dangerous information, that's going to be hard for them to tell the difference, and understand that it's not always correct, accurate information.

Host: And how well are current chatbot platforms designed to protect kids from inappropriate content or even manipulation?

Mark Connelly, Professor of Pediatrics: Unfortunately, the exponential development of these tools over the past few years, regulations just have not been able to catch up essentially. And that's starting to happen, in some, more frankly, in other countries than here to some extent.

But most of the regulations do not really speak specifically to the needs of kids. And the are often not great for protections for adult use either. And so, in short, there's just not a lot of regulation, and help with the protection of kiddos. That again, may change in the upcoming years.

But that's a major concern right now is ultimately it is frankly, left to parents on really being thoughtful about how to safeguard kiddos because there's just not a lot of regulation. And there's no requirements, for many of these companies to be thoughtful about the safeguards that they put into place for kid use, particularly chatbot Zoom.

Host: What signs should parents look out for to know if a chatbot is having a negative impact on their child?

Mark Connelly, Professor of Pediatrics: I think you can view chatbot kind of flags, if you will, as, as similar to like any other digital tool or, that you might get concerned about. So there's potentially a real risk of getting more dependent, if you will, on that becoming your companion or kinda emotional support or where you go for most information.

So just like other, if you will, addictions, it's really that they're spending most of their time on chatbots, if or on the computer or digital tools, if it's taking over for more typical social relationships. They're becoming more and more isolated, that they're referring to even these companion chatbots as things like their best friend.

Those type of things are certainly flagged to look out for that it's replacing ultimately the connection that we get from, from true bonds with same age peers and adults.

Host: So how can families set healthy boundaries around chatbot use?

Mark Connelly, Professor of Pediatrics: I think that's really the key as well. It shouldn't be all doom and gloom. I mean, ultimately these tools are really quite fascinating for their capacity to really understand patterns of language, which can be really helpful for kids that don't really understand sometimes about social cues and how to kind of communicate in different conversations. So it actually can be useful for like, rehearsing social skills. So we don't want to suggest that the best response for parents is just to make sure that their kids are never using these tools. It's just like anything else with digital tools.

 A couple things that I recommend. One is I think there's a misunderstanding to some extent, with what these tools are and are not. A lot of us might be scared of what artificial intelligence is or is not, and just don't have the knowledge necessarily to share.

So I think co-learning with kids and developing that literacy together, this is a really fantastic opportunity to learn some kind of critical thinking, and asking questions so, when is it not good use of a chatbot or what are potential problems that you see?

And really open up a conversation, around digital tool usage. Other things, for digital protections as well, again, just being particularly for younger kids, there's really no good benefit, for the most part of starting too early. So usually recommendations are trying, to delay digital exposure as long as possible, particularly not until 12, 13 or so. And I would say the same probably for chatbot use. And then, being good models as parents as well, not over reliance or overuse on chatbots or, or digital technology as well. I think it's certainly okay particularly for younger kids having a lot of supervision.

Again, the, the co-use model. It's interesting to kind of see the benefits of these type of tools. So we're using that with your kid and teens and having dialogue around that. And trying to limit some of the unsupervised use just because of some of the potential for information that comes out of these that can be harmful for the kids.

Host: Well, it's reassuring to hear that there are some pluses to chatbots.

Mark Connelly, Professor of Pediatrics: Which again, I think, the, the idea of there's lots unknown about artifical intelligence, and of course that promotes fear to some extent. So I don't want that to be the message for parents. I do think there's lots of interesting things that we don't know. And, back to one of your questions about the impact on emotional and cognitive development and social development and so forth.

Without research, we cannot say that it's all unhealthy or when isn't healthy or when is going to and for kids that are isolated, for kids that maybe even are pretty depressed and have a lot of mental health issues. It's not always to say that being on chatbots is that necessarily a bad thing.

And the perception of some companion, frankly, can be sometimes helpful, for some of those kids who stay isolated. So it's just hard to say really one way or the other if it's always going to be negative. So I think again, just being thoughtful, moderation, having good communication, all the good parenting stuff, that are generally common sense parenting strategies.

Host: Well, thank you so much for sharing your knowledge on this important new topic that a lot of us don't know too much about. We really appreciate your time.

Mark Connelly, Professor of Pediatrics: Absolutely. Thanks for having me.

Host: Again, that's Mark Connolly. To find out more, please visit children'smercy.org/parent-ish. That concludes this episode of the Parent-ish Podcast. For more parenting tips and tricks, visit us@parent-ish.org where we help you celebrate the craziness and challenges of parenthood.