Raising AI-Smart Kids – PediaCast 618

Podcast: Download

Subscribe: RSS

Show Notes

Description

Artificial intelligence is no longer science fiction… it is becoming part of everyday life! Kids are using AI to complete homework, create artwork, write stories, brainstorm ideas, and even seek advice or companionship. But what does this mean for mental health and development? Join Dr Mike and experts from our behavioral health team as we provide tips for raising AI-Smart Kids!

Guests

Dr Emre Sezgin
Principal Investigator, Center for Biobehavioral Health
Abigail Wexner Research Institute
Nationwide Children’s Hospital

Dr Ian McKay
Pediatric Psychologist
Nationwide Children’s Hospital

Dr William Leever
Pediatric Psychologist
Nationwide Children’s Hospital

Links

• Pediatric Psychology and Neuropsychology at Nationwide Children’s Hospital
• AI and Pediatric Mental Health: Joint AI Engagement for fostering Safe and Supportive AI Use
• Behavioral Health and Generative AI: a perspective on future of therapies and patient care
• The Pediatric AI Readiness Framework: bridging evidence to practice in pediatric artificial intelligence
• Attention is All You Need? When Responsiveness Short-Circuits Responsibility

Episode Transcript

[Dr Mike Patrick]
This episode of PediaCast is brought to you by Pediatric Psychology at Nationwide Children’s Hospital.

[MUSIC]

[Dr Mike Patrick]
Hello, everyone, and welcome to another episode of PediaCast. We are a pediatric podcast for moms and dads.

This is Dr. Mike coming to you from the campus of Nationwide Children’s Hospital. We’re in Columbus, Ohio. It’s episode 618.

We’re calling this one Raising AI Smart Kids. I want to welcome all of you to the program. We are so happy to have you with us.

You know, artificial intelligence is no longer something from science fiction. In fact, it is becoming part of everyday life. Kids are using AI to answer homework questions, create artwork, write stories, brainstorm ideas, and even seek advice and companionship.

But what does this rapidly changing technology mean for children’s mental health? And how can parents help their kids use AI wisely while avoiding potential pitfalls? We’re going to answer questions like these today as we separate hype from reality and share some practical guidance for families navigating this age of artificial intelligence.

Of course, in our usual PediaCast fashion, we have some terrific guests joining us in the studio to discuss the topic of today’s program. Dr. Emre Sezgin is here. He’s a principal investigator with the Center for Biobehavioral Health at the Abigail Wexner Research Institute at Nationwide Children’s Hospital.

We also have Dr. Ian McKay and Dr. William Leever. They are both pediatric psychologists at Nationwide Children’s Hospital. Before we get to them, I do want to remind you the information presented in our podcast is for general educational purposes only.

We do not diagnose medical conditions or formulate treatment plans for specific individuals. So, if you’re concerned about your child’s health, be sure to call your health care provider. Also, your use of this audio program is subject to the PediaCast Terms of Use Agreement, which you can find at pediacast.org.

So, let’s take a quick break. We’ll get our experts settled into the studio, and then we will be back to talk about raising kids in the age of AI. It’s coming up right after this.

[MUSIC]

[Dr Mike Patrick]
Joining us today are three outstanding experts who bring unique perspectives to our conversation. Dr. Emre Sezgin is principal investigator in the Center for Biobehavioral Health at the Abigail Wexner Research Institute at Nationwide Children’s Hospital. Dr. Emre Sezgin is here. He’s a principal investigator with the Center for Biobehavioral Health at Nationwide Children’s Hospital and he is an assistant professor of pediatrics at the Ohio State University College of Medicine. His research explores the intersection of digital health, artificial intelligence, and pediatric care.

Also with us is Dr. Ian McKay, pediatric clinical psychologist at Nationwide Children’s and assistant professor of psychiatry at Ohio State. His clinical and research interests include child and adolescent mental health, digital technology, and helping families navigate an increasingly connected world.

And we are pleased to welcome Dr. William Leever. He is a pediatric psychologist at Nationwide Children’s and assistant professor of pediatrics at Ohio State. His work focuses on children’s emotional wellbeing, healthy development, and the thoughtful integration of technology into mental health care.

Together, our guests will help us understand what AI means for today’s children and teenagers and how parents can help their families use these powerful tools safely, thoughtfully, and responsibly. Before we dive in, let’s offer a warm PediaCast welcome to our guests, Dr. Emre Sezgin, Dr. Ian McKay, and Dr. William Leever. We are so happy to have all of you with us today.

And I wanted to start with you, Emre, just with the basics. You know, what exactly is artificial intelligence and what do parents sort of understand, you know, AI to be when they hear folks talking about it?

[Dr Emre Sezgin]
Dr. Mike, thank you for having us, first of all. AI in the fundamental terms is a computer program or a machine that built to do tasks normally require human thinking, such as recognizing patterns, speech, solving problems, completing tasks, and learning from examples. But nowadays we are using AI as an umbrella term, which kind of in our current ecosystem, and we actually lead with AI in different forms for many years, like recommender systems, auto-complete when you’re typing on your keyboard, navigation systems, fraud detection, and this has been already using these systems, AI, in a form.

What has changed dramatically over the time is the last few years we have seen generative AI tools become more accessible directly to us. And nowadays a person can now talk using their natural language to an AI system, ask questions and create some things and ask for some advice. And we are living in a new era of an AI.

[Dr Mike Patrick]
Yeah, absolutely. Ian, how are children and teenagers actually using AI? I mean, I know for myself what I’m doing with it.

Are kids also using AI quite a bit? And what sort of tasks are they using it for?

[Dr Ian McKay]
Yeah, well, again, thank you so much for having us, Dr. Mike, and it’s a phenomenal question. I think one of the biggest misconceptions that I often hear is that kids are only using AI to cheat on their homework. And I think the reality is a lot broader than that.

We’re seeing AI become integrated into many different aspects of kids’ life and how they’re using can vary considerably across development. We do know that kids from a developmental standpoint, their social needs, their cognitive abilities, as well as their independence are evolving over time. And so naturally we’d expect how they’re using these tools to evolve over time as well.

If I sort of break this into three categories and start just with the younger kiddos, I think their use is much more invisible. The AI systems are often integrated into some voice assistance, also through apps or through various browsers that kiddos may be sort of using and even recommendation algorithms. And it’s important to note that kids at this age are still sort of like learning to distinguish what is real and what is not real or what is fantasy.

And so, as a result of that, the adult involvement I would say is quite important here and we’ll talk more about that later today. But we’ll also mention that there is still a lot of importance of like internet guidance and oversight that is recommended for kiddos this age that are using artificial intelligence, especially since their use is more invisible. For that second phase, we think about like elementary school kiddos and rather than it being more invisible, their use is a lot more intentional.

I’ve seen kids use this to ask questions about creating a story, creating images, asking for help on a difficult math assignment. And so, their use again can differ depending on age and it becomes really at this stage more of a tool for exploration and for learning. And then when we think about adolescence, this third stage, it shifts from being like less of a novelty tool to more of a productivity tool.

So, we’re seeing kids use this for things like generating a resume or for getting help on a homework assignment or generating ideas for a particular essay that they might be tasked with completing or asking questions about careers. And so, what I will point out is that it’s important to consider that it’s not really about whether a kid is using AI or not, but rather how are they using it and the role it’s playing and if it’s supporting healthy development.

[Dr Mike Patrick]
Yeah, yeah, absolutely. Will, I wanna bring you into the conversation. You know, when we think about AI and the possibilities, like there’s excitement on what these things can do and then there’s also the concern.

We hear from the tech engineers and the CEOs of the company, like putting out warnings, like it’s gonna take over the world and humanity is at risk. What do you see as sort of the real risks and the biggest opportunities as well for families and kids in particular?

[Dr William Leever]
Yeah, well, again, just wanna echo my colleagues. Thanks for having us on. I think this is a hugely important topic.

I like the way you frame this question because it really is both. I mean, there’s the existential risks and opportunities, but we’ll talk a little bit smaller here around for kids. So, I see, I’m gonna simplify this to two big areas.

There’s endless opportunity, endless risk, but there’s also these really great windows that we can look into to simplify this. So first, the opportunity really falls into personalization. AI tools are incredible at really meeting people where they are.

And so, it opens up windows for practice, creativity, learning, and all these other areas where you can really dig in and explore topics or things that you may not have had access to. On the risk side, there are things that AI can add to the problem. Like there are these big risks that make headlines.

However, I like to focus on a smaller, subtler risk, which is replacement. And AI can replace a lot of things in a kid’s life if we’re not careful. So, I have three big areas that I look for.

One is relationships. I’m big on social connection. And as AI gets better, especially generative artificial intelligence, and we move more into agentic intelligence, which Emery will talk about later, we see this ability to basically be a frictionless friend and really replace those family and friend relationships that are a part of growing up.

Second piece is effort. If you use AI passively, just give me the answer type of approach, it really can reduce the amount of learning that we’re doing. And then the last one is discomfort.

And this may sound weird, but I think that kids have to feel uncomfortable growing up. You have to go through awkward relationships and fearful situations. And AI, again, can make things really frictionless.

It can take away that and really put us in a place where we’re going there for our comfort and for all of our answers. So those are the areas where I see is the biggest risk right now.

[Dr Mike Patrick]
You know, when we think about a new technology, there’s always sort of fear and, you know, expectations that things aren’t gonna necessarily go well. Oftentimes then when we get wherever we’re going and we take a look back, you know, we can see the benefits of the challenges that we took. What do you think the biggest thing right now that parents can do to sort of, you know, make it easier in the future for them to look back and say, okay, everything went all right.

Like, you know, as parents, we’re kind of guiding our kids on journeys. And in this case, that journey is the use of AI.

[Dr William Leever]
We wrote a paper on this, and I think this is a really important topic. And one of the ways I describe it is just learning together. So more formally, we call this joint engagement.

But the way I think about this is you have to have these conversations. Kids are already doing this. I mean, if you look at the statistics, anywhere from 65 to 90% of kids have interacted with some sort of generative artificial intelligence, those numbers are not accurate right now because it changes every month.

But what that means for parents is you have to have these conversations, and you have to take this stuff seriously. It’s not a passing fad. It’s not just something that’s going to blow over.

And we know with the danger from social media and machine learning algorithms, this stuff is basically a large experiment with our children. And we need to take the reins and actually start to be proactive with how we deal with it. So, learning together, have a conversation, sit down.

One of my favorite things to do for those who are having this conversation with older kids is have them teach you what they’re already doing, what they already know. Have them teach you what they know about AI, because I know a lot of parents will come in and say, I don’t know anything about this stuff. But still have the conversation, still learn together, even if it’s your kid teaching you.

But at that point that feeling powerless and feeling helpless and feeling like this is a wave that’s washing over us, we always have control and agency of how we deal with this. It’s just another parenting thing like we deal with all the time. And so not feeling too overwhelmed and just being able to engage.

[Dr Mike Patrick]
Emre, you guys developed a framework for families in dealing with AI and incorporating it in a healthy way into families. Why did you feel families needed a framework instead of just general advice?

[Dr Emre Sezgin]
Yeah, that’s a good question. Frameworks have been historically helpful for everyone to understand and conceptualize new ideas and terms and new technologies. When we look back and try to explain new technologies to our families and patients, let’s think about like mobile applications, how we can use this to kind of like help you and then keep you in track of like medications and other things that we need to keep track.

I mean, kind of like in the same situation here, but right now it’s really harder to conceptualize because the AI using your natural language, there are kind of endless opportunities you can achieve, right? And there’s those endless risks you can, or you cannot even think of. So, we like to help families and our children conceptualize these technologies better and starting with maybe baby steps, like step-by-step, what you can do, what you cannot do or what you shouldn’t do.

So that’s why we come up with some ideas of this, like Will mentioned, joint media engagement. And before we had some work around like how we can guide not only families, but also the decision makers and the institutions to better guide in the broader communities. And that has been historically really essential and effective to understand and communicate these technologies and being safe and responsible.

You defining the end users, the controlling bodies, who’s accountable and liable. And that kind of makes everyone being aware of what you can do and what you shouldn’t do or what you shouldn’t let people do with these technologies. And that kind of been very helpful for us to develop and work on.

[Dr Mike Patrick]
And Ian, what does healthy AI use look like within a family? So, if we’re thinking about this framework and making critical decisions, teaching our kids to make critical decisions on different aspects of AI, what is healthy AI use?

[Dr Ian McKay]
Yeah, another really important question I appreciate you asking. And I think that how we define healthy AI use can of course vary depending on who you’re asking, but also the context in which this question may apply. I think today I’ll just briefly discuss two areas where I think this is really relevant, both in the home and within the educational environment.

And we’ll also point out that as psychologists, we do know that kids learn best when they’re actively engaged in the learning process. And so, I like to think of AI as another educational tool that it works best when it sort of scaffolds learning rather than replacing it. So, if we think about the school setting first or educational setting, that kids can use AI in a healthy way to help maybe explore a concept in a different way or to have it explain something that might be difficult for them to possibly generate additional quiz questions to help reinforce knowledge that they are learning, to help them to create a study guide or to acquire a new language.

And in each of these examples, the child is really doing the thinking, and AI is supporting the learning process rather than doing it for them. If we think about the home environment, I think healthy AI use can be a really great opportunity for shared learning. As others thus far have pointed out, there can be opportunities to ask questions together and explore what the answers may be and whether it should be verified elsewhere, whether that information seems accurate.

And the American Academy of Pediatrics also recommends that this kind of modeling is really important for helping individuals or kiddos specifically to acquire these critical thinking skills. And I will lastly point out that, I don’t think that healthy technology use has really ever been about the technology itself, right? It’s not that an iPhone is problematic, but how one uses it can become problematic.

It’s really more about balance. And we know that kids developmentally still need opportunities for things like face-to-face relationships as Will had pointed out or physical activity or creativity and unstructured play. And there have been a lot of concerns raised which are valid by some scientists that have been concerned that the technology could possibly crowd out opportunities that we know are really essential for healthy adolescent development.

Things like learning to tolerate distress or learning to navigate social relationships or to be uncomfortable as Dr. Leever pointed out or forming an identity. So, to sort of sum that up, I don’t think the goal is really to avoid AI, but rather to help kids learn like when it might add value to their lives, when it might be important to seek the guidance of a trusted adult or when to think independently.

[Dr Mike Patrick]
And William, a term that families may hear is AI sick of fancy. And I actually have trouble pronouncing this word and I had to like do it phonetically so I was saying it correctly. What is AI sick of fancy and why should parents be aware of this?

[Dr William Leever]
Yeah, well, this is just a big word for a very common concept. So, this is essentially a fancy way to say flattery or overly affirming of our own opinion. So, AI and sick of fancy really go about telling you what you wanna hear.

And if you’ve ever used a chat bot, it’s gonna, whatever you say, it’s gonna be the best thing it’s ever heard. And there are two factors with this. So, the first factor is a little bit of sick of fancy is actually a good thing.

So, there was a study and this is counterintuitive, but there was a study that if you approach someone with a little bit of validation for their opinion and you approach them gently, this AI study, they actually made better decisions after they were approached in that way because you don’t put up your defenses. The problem is when you turn that volume all the way up, it really helps with engagement, which is the ultimate goal of a lot of these commercial chat bots is to keep you using them. And people really like it, they trust it, but it makes you more entrenched in your own opinion.

So, there was a study in science where they had patients rate their reactions to a conflict. And the more sick of fancy that was introduced into their opinion, validating their opinion, the more they left that without taking responsibility for any role in the conflict. We use a tool in clinical work called motivational interviewing, and that involves a lot of validation.

You really wanna make people feel heard, but you wanna validate the things that are moving them in a healthy direction. So, you wanna kind of catch on to that thing that’s really sending them in the right way, like the effort they’re putting into resolving a conflict. And if you validate or you’re sick of fancy in everything, people will like it because people love to hear that they’re right, but it won’t help them make good decisions or get the truth.

[Dr Mike Patrick]
Yeah, it’s one of those things where when you hear something over and over again, you start to suspect it. And maybe it’s not being genuine or authentic if an AI system can be those things. But as humans, that’s what we want when we engage.

And so, I can see where like a little flattery is fantastic and makes us feel better about ourselves and wanna keep using the system. But too much, it does start to feel a little strange, non-human, if that makes sense.

[Dr William Leever]
This is especially important for those with neurodivergence too. There’s a lot of risk that comes with excessive sickness around things like if you’re hallucinating or you’re suicidal or you have these other major issues in your life. So that’s a danger to watch out for parents as well.

[Dr Mike Patrick]
Yeah, and we do hear about stories of this happening. I get the sense that it’s not often because I feel like it’s sensational enough that we would hear in a country of over 300 million people, we hear an occasional case of this, but it seems to be the exception, not the rule, but still, something that parents need to be aware of.

[Dr William Leever]
Very good point, yes, very true.

[Dr Mike Patrick]
Yeah, and I’m talking about when a kid has suicidal thoughts and they go to a chatbot and start speaking and then the chatbot is encouraging them. So that’s where really, we wanna be aware of what our kids are talking about, not only with their friends, but with AI chatbots as well. Another issue that often comes up is this idea of hallucination.

And I think that kind of goes along with the flattery thing because the AI wants to be able to provide what you’re asking it. And sometimes if it doesn’t know, it just makes something up and we call that hallucination. Ian, can you speak to that and how can we teach our kids to recognize when an AI is hallucinating, especially if it’s a topic that we don’t have a lot of foundational knowledge in?

[Dr Ian McKay]
Yeah, so I think one of the simplest ways I can explain what an AI hallucination is essentially that the AI doesn’t know when it doesn’t know something. And unlike a teacher or a colleague, a physician or a peer who may acknowledge that they do not know something, the AI systems in which we’re speaking about today, more specifically large language models are designed to generate a response regardless. And they essentially, the most simplistic way I can describe this is that they are producing a response that is most statistically probable and not necessarily the most truthful answer.

Sometimes these responses can be excellent, which we do see a lot of the time, but sometimes they are not excellent and can have made up citations, which we’ve seen in the media, incorrect statistics, or just entirely made-up facts. The challenge here is that regardless of if they’re accurate or not, they present this information with the same level of confidence. And so, this can be concerning.

People tend to maybe believe information that is generated by a machine is more objective. You see this in pop culture, if someone shows their phone and they say, look, this is what the internet says without really understanding if that information is correct or being able to sort of discern whether it’s valid or not. And this can make the hallucinations more persuasive than they should be.

And this is really important, especially when we think of those that are under 18 years of age who are still developing those skills needed to evaluate information more critically. And so, I do think that this is not something that’s really specific and unique to AI, I will call out. We’ve been teaching kids for a long time in school to be able to think critically about information they encounter, if it’s information on social media, if it’s information that they encounter on various webpages, this just really adds another source for which to sort of practice and test these skills.

So, it’s just, again, important to note that confidence isn’t really the same thing as accuracy. And Dr. Mike, I’m happy to speak as well to the piece about how do we help adolescents to recognize this? I do think that that may be a question that we’re tackling later, but I’m happy to discuss that as well.

[Dr Mike Patrick]
Yeah, let’s address that.

[Dr Ian McKay]
Sure. And so, sort of how do we help kids and parents to better recognize AI hallucinations is I think teaching them that AI is not necessarily bad or shouldn’t be distrusted, but rather we should teach kids healthy skepticism, which we know to be an important developmental skill to acquire. I think back to my school-aged years where we were sort of tasked with going to the library and trying to find peer-reviewed articles and to suss out what might be inaccurate or what may not be based in science or in fact.

And so important for kids to kind of question, like, is this information something that I would maybe trust related to my own health or related to a topic that’s very important to me or a school assignment? And these questions can help kids to become healthy consumers of information and think also parents can play a big role in this as well, where they can potentially have an opportunity to practice shared learning. I think Dr. Seskin pointed this out that we can ask questions and have families kind of practice whether those answers are correct. Maybe they have the answer initially and then they’re testing whether the AI system is generating the correct response. And then to learn that verification matters. There was one study I’ll point out that was I think really relevant here as well, where they had provided one group of subjects with some preemptive knowledge that these systems can hallucinate.

And in that group of individuals, they were more likely to catch the AI hallucinations when compared to those who were not given that preemptive information. And this is why I think we’re seeing AI literacy being built in K through 12 curriculums for this very reason. So, there’s a lot of benefit of those preemptive conversations as well.

[Dr Mike Patrick]
Yeah, yeah, absolutely. And I love that schools are incorporating AI literacy and like, how do you use it? How do you think critically to make sure that you’re getting the best information?

And I think part of that education is giving the AI that you’re using enough context of what it is you’re asking it. I have used it myself just for things around the house, like, hey, help me figure out how to do this. And when I used to watch YouTube videos to teach me how to do something, now often I just have the AI sort of guide me through.

But the more that you do that, the more that you do figure out like it needs, you add another data point and it’s like, oh, wait, no, now my advice is completely different. And so, I think just teaching good prompting and how to really put what you’re asking the AI to do into context is really an important part of that training and something that we can do as parents at home to kind of reinforce the AI literacy, the education that kids are getting in school. Would you agree with that?

[Dr Ian McKay]
Yeah, I think it’s a beautiful point that prompting is so essential and can really produce such variable responses depending on what is being asked. And so, I’ve sat with kids before who may be struggling with a particular area of school and rather than having the tool kind of do the work for them, we’ve sort of talked about ways we can prompt these tools collaboratively to help them to learn what is required more independently and to practice that. And so I think it’s a great point you bring up that we could really benefit from helping kids to prompt and doing that co-prompting together, part of that JME framework that we spoke to earlier can be really essential for the opportunity to learn together about these tools.

[Dr Mike Patrick]
Yeah, yeah, for sure. William, I wanted to shift to mental health and specifically AI used in mental health care. Is that being done?

Are we seeing more and more where folks maybe go to AI instead of a therapist? And what does the evidence tell us about the use of AI in mental health care?

[Dr William Leever]
Yeah, well, I’m actually gonna take a step back because AI is being used all over the place in mental health care. And in a lot of ways, it’s really great. So, there’s four big areas that I’ll start with.

So, the first one is administrative. The second one is assessment and prediction. The third one is training and supervision.

And then the fourth one is that direct intervention where we really think about like people going to a chatbot for advice like they would a therapist or something like that. In the beginning, the administrative part, AI could be really revolutionary here. So, Emery and his team and a lot of people were working on some really great ways to automate or reduce the amount of administrative burden that mental health providers have to deal with.

And I think this is just gonna be an amazing improvement in the relationship. So, coming back to that social connection idea, I think eventually generative artificial intelligence and agentic AI can really actually make the healthcare experience more interpersonal, more social. The second piece where it’s really amazing is with assessment and prediction.

We’ve got some amazing people at the hospital here working on this. So, for example, Dr. Eric Youngstrom and his team are using artificial intelligence to improve our ability to make diagnoses, mental health diagnoses, and really review people’s patients’ records and things like that. And I’m very excited about that.

The third piece is actually an area that I’m personally interested in, which is supervision and training, where instead of people getting practice on real people, we can actually do a lot of this with generative artificial intelligence. And so, we’ve worked on some projects where we help with the training and the role-playing practice side. The last piece is by far the most dangerous right now.

It also has a lot of potential, so I don’t wanna be all negative around this. But the direct intervention piece, a lot of people are going to, and young people are going to, chatbots for mental health advice. And I think that’s because it feels private.

It feels like it’s not giving your information out there. That’s really misleading, because all of that information is actually not private whatsoever. The advice can also be overly sycophantic, like we talked about.

It’s gotten better, and it actually has a lot of potential, but right now there’s just no guardrails. There’s nothing that’s really keeping it on track. There are some places, like there’s some researchers at Dartmouth who’ve looked at using generative chatbots for actual therapy with a real research infrastructure.

So, I’m optimistic that this will be something we can use in the future. But right now, I would really argue against an unregulated use of chatbots for mental health. You really need that human in the loop, and it’s important for you to see a professional for this.

So, I would say it’s a mixed bag, opportunities and risks, just like we talked about before.

[Dr Mike Patrick]
Yeah, yeah, for sure. You know, access to mental health care can sometimes be difficult, and especially in rural areas, places where there’s underserved, so there aren’t as many therapists as possible, there aren’t psychologists maybe in particular areas. And so, when someone gets a referral to see a mental health professional, if there’s not an emergency, like I’m actively suicidal or am engaging in self-harm, sometimes there’s quite a wait to be able to see a mental health professional.

And then that makes going to a chatbot even more appealing that, hey, I can talk to someone about this right away. Emre, can you talk a little bit about access and how sort of we can use this technology to help people, but also make sure that they’re, you know, getting human professional help at the same time?

[Dr Emre Sezgin]
Yeah, definitely. This topic has been one of the main research areas for me for the last couple of years. I’ve been trying to investigate with my colleagues here.

Access, when you think about AI, at the first glance, we think that, hey, okay, this is kind of leveling the ground. Now I can get the information in my own language, and I can just learn more about this specific topic or a health condition or the response that I received from my physician that I don’t understand, I can kind of make more understandable. So, we have seen many, many examples that it helps improve the access to the health information.

And I’ve been working on this topic a bit more like how we can kind of going beyond that, because we know this now, the AI, right now we are using trained on the data that has been available on the internet. And when we think about this, it basically gives you the idea of like how this data being collected and being used. Internet requires internet connection and data plans.

And that is kind of gives you, it creates a cohort that already has some certain privileges, the access to the information. And we developed our AI tools around this information. And now we are trying to use it with the populations we have never had a chance to collect data or interact with in rural or Appalachian areas or underserved populations.

So that kind of creates a bit of kind of imbalance. The tool tries to suggest, when we try, we have some research projects around this that we try to use AI to serve an immigrant population. Like we build personas and cases or scenarios that pretend as like a family from, assume that immigrated or a refugee family trying to help find some information.

It’s most of the time, it fails to provide culturally appropriate information and responses. So that’s when we started to think about how we can improve these technologies to better serve these populations. And that’s why when we think about this basic needs, like food assistance, transportation or childcare, housing support for the communities that has historically been underserved, we decided to take a step forward to use and make it more accessible, make the resources more accessible through this AI tools.

And we have been working on one potential use case for conversation AI is reducing this information navigation burden for the families to access these appropriate resources. And what this means for the communities that we are talking about, it can be like immigrants, refugees or also populations rural and Appalachian areas. It gives you a low-cost opportunity to access this information.

You don’t need to call a service, or you don’t need to connect with the hospital’s social worker information centers. You can just ask these technologies and find the resources. And it can give you the opportunity to ask these questions 24 seven.

You don’t need to worry about the language barriers. You don’t need to worry about your literacy level. It can be health literacy or kind of like internal literacy.

The information can be translated or turned into more understandable kind of level of information. And support for specialized professions could be a bit more accessible in terms of like understanding the information in the specific use cases. For example, we’ve been talking about and doing some research about the oncology space.

And there’s some sort of questions like families are asking about, let’s say chemotherapy or going back to school and some of the side effects of certain medications. And AI, it has been able to interpret and communicate this information like clearly that we may not be able to access before through our internet search engines. I’m not saying here it’s gonna replace like the professional’s role, but at least it can give you some of a bridge.

Like when you cannot access this professional services, it can be a remedy to interpret and understand this information that you need and at the time that you really need to interpret this information. And we can see in our research and the evidence shows that it can help families navigate the fragmented systems mostly like in healthcare.

[Dr Mike Patrick]
Yeah, gosh, when we think about healthcare and asking questions and looking up answers on the internet traditionally, as we’ve done over the last several decades, this has really changed things in terms of drilling down and getting what you want. But you really have to, just like anything, it takes practice to ask it in the right way to get a good answer. And that again, where that AI literacy and fluency is just so important, not only for kids to get at school, but also with their families too.

Ian, I wanna focus in on chatbots here for a moment. You know, there are times when kids are lonely or they’re stressed and need to talk to someone, and the chatbot may be the only option. Should parents be concerned about that?

Is it all bad or is there also some good that can come out, especially as we’re talking about being stressed and lonely?

[Dr Ian McKay]
Yeah, we’re seeing this certainly become a really hot topic in the area of research and is something that continues to be explored that we don’t have a ton of data on at this current moment. I will say that the AI chatbots are extremely accessible. They’re available 24 seven for folks to use.

They’re often free, depending on which version people are using, though the free versions tend to be for folks, they may perceive them to be sufficient in this particular regard. I will also say that they help reduce a lot of the burdens that we saw in access to care challenges, especially amidst this mental health crisis that emerged a few years ago, where we saw significant wait times, barriers to services, as well as just challenges maybe engaging with the particular services that folks needed at that point in time. And so, these chatbots really help to solve that challenge for folks.

They also tend to be extremely validating as Dr. Lieber pointed out earlier, which can be really useful and beneficial for some individuals. But what’s important to note is that they can feel very human-like, but they are not. They do not have a conscious.

They are not able to do things that trained mental health providers go to school, get a lot of training for, a lot of supervision with, and experience for. And they are, for example, not able to take all of the individual’s historical context into account when making a diagnosis or when providing support. They are not able to consider developmental history or maybe family history of mental health concerns.

And so, they are not able to provide the same level of support that we would expect from a human and a trained mental health clinician. And so, I will say that these tools are here. People are using them, but there is some data to point out that about 61% of users that are using these tools for mental health reasons are not disclosing them to their providers, which is certainly a concern.

So, this just sort of helps me, and I think some of our colleagues here think about better efforts to screen for AI use when we’re working with kiddos, asking them about their use, normalizing that, trying to break down some of the barriers to disclosing their use. And then also engaging in conversations once that sort of door has been opened around how they can safely use these tools. If they’re showing their chat histories with mental health professionals or parents, and that I think like anything, it’s really more about balance.

There can be some benefits of maybe low-level needs, not enough that we can confidently speak to that at this point in time. And there’s a lot of benefit of keeping that provider and human in the loop and then discussing how these tools can maybe be used outside of sessions in more adaptive and productive ways.

[Dr Mike Patrick]
Yeah, absolutely. Emery, as folks use chatbots, especially when they’re stressed or lonely, they can start to really have a relationship with the chatbot who seems so real, which does not seem, I feel like now we have crossed into unhealthy when that happens, like when we have feelings of love or strong relationship feelings for a chatbot. What are some warning signs that parents should be on the lookout for that that might be happening?

And if there is a concern that feelings are developing for the chatbot, what should parents do?

[Dr Emre Sezgin]
That’s a very good question. And yes, we have been seeing the news and the people showing interest to this, to having this close relationship with these agents, these machines. It’ll also remind us of some of the movies that we saw like a couple of years ago, like Her.

It’s kind of getting a bit scary. And as parents, I also have two young kids and I’m thinking about the future for them and how we can guide our kids. That really goes back to our human relationship.

Just we need to remember these AI tools are like any other technologies. Maybe they have more capabilities, but they are kind of basically built as a tool to help and serve humans. And when we are working with those tools, they’ve been designed to gain your trust, gain your interest.

When you think about like the four mobile apps that they’re kind of like social media tools, they’re kind of designed to engage with you, right? The more you stay on the page, the more it is profitable and good for the services when you think about like this commercial tools. And AI kind of uses natural language to do that.

It confirms all your actions; it sends you some emojis to build this trust and connection. And thinking about our loneliness epidemic right now, and we have some survey results from our underserved cohorts that we did some research. They also kind of almost like up to 60% reported loneliness.

And that makes these accessible tools say kind of like something they can seek some companionship. And unfortunately, since some of the specific AI tools built towards that concept, like, hey, this can be your companion, this can be your trusted partners, you can share everything. Things can really go out of hand.

And we already see that. Unfortunately, we have seen teenagers using this technology, sharing some intimate information with this AI tools that they don’t share with their parents and with their friends. And may lead to some sort of like mental breakdown and suicidal attempts as we have seen in the last couple of years.

My colleagues here can talk about this better because I don’t have like the expertise in the mental health, but as an observer, and as a parent, I can say these technologies are becoming more concerning in terms of when they build something around this companionship. Yes, we see some users develop some strong attachments. What we need to do is kind of right now having this open conversation.

This has been like our whole topic in this podcast. Like we should, as parents, we need to have this open conversation with our kids. Asking their opinions and asking them how they use these technologies, maybe in their own time, we may not be able to observe that, and how we can guide them to have more, healthier conversation with this technology, healthier relationship with this technology, instead of kind of investing too much time and connection with this non-authentic or simulated tools.

And as we can see right now with the robotics, I just need to take us to a couple of years future maybe. We see like low-cost robotics and people are trying to create this humanoid devices that can also show this kind of human expressions. Like you can see a screen with a face smiling at you, and that may also increase the risk of this.

So, we should now actually, we shouldn’t think about this is going to be, hey, this is going to be like something in the future, like it’s kind of science fiction. Now we need to just sit down and think and talk with our kids and try to kind of guide them how they need to perceive these technologies in the future for their, as a tool and for their kind of like relationship and where they need to stop and not go further with their communications and kind of connection with the technology.

[Dr Mike Patrick]
Yeah, absolutely. And it’s one of those things where the more that we have our finger on the pulse of our kids’ lives and we’re talking to them in an open way and letting them know they can talk to us about anything that’s concerning them and starting that at a young age so that those kinds of conversations are normalized as things get more and more complicated, which I think is going to happen. You know, one of you said that AI is not going anywhere and it’s not.

And the forms of AI that we have right now are the dumbest and least complex that we’re going to be dealing with. And then Emre, when you mentioned, oh yeah, these humanoid robots, and now we think about the AI that we’re seeing today attached to those, like this is just going to get more and more complicated as we go. And so, I think, you know, families talking about these things right out of the gate is just going to be so important.

Will, I wanted to end with you as we wrap up, you know, as this does all become part of everyday life, what are some important things that parents can do to prepare our kids for the future? If you had to sum all of this up into a few really important points, what do you have for us? One word I would use for this is efficacy.

[Dr William Leever]
And that’s a little bit clinical, but not too much. It’s basically feeling like you can be effective with this stuff. And I think it’s so important right now when people are feeling a little bit hopeless or powerless to deal with the AI influx.

You have the ability to manage this. And so, I want parents to feel like they have parental efficacy. You can do things about this.

I would ask questions like we’ve been talking about. I would set boundaries. It’s completely appropriate really throughout the age span to set boundaries around AI.

Shouldn’t be up late at night talking to an AI chatbot when you should be sleeping. I think you should try and learn about it yourself. I think that an important part of this is integrating it into healthy habits.

In a way, it’s kind of like healthy AI hygiene. Like you don’t share all your personal information with an AI chatbot. You don’t allow it to answer all your questions for you.

You do your own homework. And so, as we go through this, I want people to take a lesson of like, I got this, I can handle this. And there’s an efficacy or effectiveness component to all of this.

With that being said, I really wanna keep the human connection at the center of all of this too. And I think that that’s an important piece to feeling effective is getting support and talking to other parents or professionals about this. And I think that can really help with taking down some of the existential drought around AI.

[Dr Mike Patrick]
Yeah, absolutely. This has been a fascinating conversation. And it’s one of those topics that I think we could have you guys back in a year and it’s a different conversation.

And I think what’s five years from now? I don’t even know. So, but it has been terrific and we really do appreciate you guys stopping by.

If folks head over to the show notes for this episode, PediaCast 618 over at pediacast.org, we are gonna have some resources for you, including some great journal articles that really dive into this a little bit deeper. So, if you wanna get more sciency, we are gonna have great resources, including a paper that the three of you put out in NPJ Mental Health Research called AI and Pediatric Mental Health, Joint AI Engagement for Fostering Safe and Supportive AI Use. And we’ll put a link to your article in the show notes as well, along with those other resources.

So once again, thank you to Dr. Emre Sezgin, Principal Investigator with the Center for Biobehavioral Health. Dr. Ian McKay and Dr. William Leever, both pediatric psychologists at Nationwide Children’s Hospital. Thank you all so much for stopping by today.

[Dr Emre Sezgin]
Thank you very much.

[Dr Ian McKay]
Thank you.

[Dr William Leever]
Yep, thanks.

[MUSIC]

[Dr Mike Patrick]
We are back with just enough time to say thanks, once again, to all of you for taking time out of your day and making PediaCast a part of it. We really do appreciate your support. Also, thanks again to our guests this week, Dr. Emre Sezgin, Principal Investigator with the Center for Biobehavioral Health at the Abigail Wexner Research Institute at Nationwide Children’s Hospital. Also Dr. Ian McKay and Dr. William Leever, both pediatric psychologists at Nationwide Children’s Hospital. Don’t forget, you can find us wherever podcasts are found. There might be an easier way for you to subscribe and listen.

We’re in the Apple Podcast app, Spotify, iHeartRadio, Amazon Music, Audible, YouTube, and most other podcast apps for iOS and Android. Our landing site is pediacast.org. You’ll find our entire archive of past programs there, along with show notes for each of the episodes, our terms of use agreement, and the handy contact page if you would like to suggest a future topic for the program.

Reviews are also helpful wherever you get your podcasts. We always appreciate when you share your thoughts about the show, and we love connecting with you on social media. You’ll find us on Facebook, Instagram threads, LinkedIn X, and Blue Sky.

Simply search for PediaCast. Don’t forget about our sibling podcast. That would be PediaCast CME.

It is similar to this program. We do turn the science up a few notches and offer free continuing medical education credit for those who listen, and it’s category one credit, and not only for doctors, but also nurse practitioners, physician assistants, nurses, pharmacists, psychologists, social workers, and dentists. And it’s because Nationwide Children’s is jointly accredited by all of those professional organizations that we can offer the credits you need to fulfill your state’s continuing medical education requirements.

Shows and details are available at the landing site for that program, pediacastcme.org. You can also listen wherever podcasts are found. Simply search for PediaCast CME.

And then we also have a faculty development podcast from the Center for Faculty Advancement, Mentoring and Engagement at the Ohio State University College of Medicine. It’s a podcast that I host, and if you are a teacher in academic medicine or a faculty member in any of the health sciences, then this is a podcast for you. And you can find FAMEcast at famecast.org and wherever podcasts are found by searching for FAMEcast.

Thanks again for stopping by, and until next time, this is Dr. Mike saying stay safe, stay healthy, and stay involved with your kids. So long, everybody.

[MUSIC]