Dr. Jeremy Sharp (00:12)
Many of y’all know that I have been using TherapyNotes as our practice EHR for over 10 years now. I’ve looked at others and I just keep coming back to TherapyNotes because they do it all. If you’re interested in an EHR for your practice, you can get two free months of TherapyNotes by going to thetestingpsychologist.com slash TherapyNotes and enter the code testing.
Dr. Jeremy Sharp (00:36)
Hey folks, I am really glad to have NovoPsych Psychometric sponsoring the show. If you do structured assessment work, then you will likely love NovoPsych. NovoPsych brings 150 plus standardized measures into one platform. What I particularly like is the extra layer of psychometric interpretation. So it helps you understand what scores actually mean. So the results are easier to communicate. If you are interested in high quality measures for personality, disability, ADHD, or autism,
You can try NovoPsych with a 15 day free trial via the link in the show notes, is novopsych.com slash testing psychologist. That’s N-O-V-O-P-S-Y-C-H.com slash testing psychologist.
Dr. Jeremy Sharp (01:21)
Hey folks, welcome back to the podcast. I’m talking about something today that is really engaging right now. It’s another AI episode, but this time it’s not related to the work that we do necessarily, at least in the assessment realm specifically. We’re not talking about report writing. We’re not talking about tools that can help your work. I’m talking about how to discuss AI with kids. So this episode is derived from
presentation that I did recently, had the opportunity to talk with a fantastic group of folks about AI in education and AI with kids and how to approach this, both from a parenting standpoint, but also an educator standpoint and looking at the different layers of how our kids might be using AI now and in the future. So we’re going to talk about general overview of AI with kids and
where we are, some statistics, talking about what I consider to be the real threat, I suppose, of AI use in kids or opportunity is another way to think of it. We talk about specific strategies to help kids use AI responsibly in a general sense. And then toward the end, I’m going to talk about specific strategies with neurodevelopmental disorders or kids who may have things like ADHD or
spectrum disorder or learning disorders. So if you have kids or certainly if you work with kids, this is, I think, a helpful episode, as always, with some pretty concrete tips to take away. Before we get to the episode, I always am like in the zone here of mentioning the Crafted Practice Retreat that’s coming up this summer in Fort Collins, Colorado. Yes, it is expensive, but it’s all inclusive. The value is there.
It’s four days of small group business coaching, time to relax, time to work on your business. Most importantly, probably time to build some relationships and get support that I don’t think many of us practice owners have in our day to day lives. So check it out. It’s the testingpsychologist.com slash crafted practice. And you can set up a introductory call to.
Figure out if it’s a good fit. I’ll talk you through it and answer all your questions. I know it’s a big commitment. So I want to make sure that you are in the right place as you decide to commit. But we’d love to have you. We have a few spots left and it’s a great event. So with that, let’s jump to the conversation about how to talk to AI.
So with that, let’s jump to this conversation about how to talk to kids about AI.
Dr. Jeremy Sharp (04:06)
All right, everybody, we are back. And as usual, I’m going to jump right to it. So like I said, this topic came from a talk that I gave recently, which is a lot of fun, to be honest. But this is an interesting talk to deliver because I was coming at it not just from the place of being a licensed psychologist who works with kids all the time or the co-owner of a tech company, an AI company that helps with report writing. That’s Reverb, of course.
But I was primarily coming at it from the perspective of being a parent in this world, because I think I’ve mentioned on the podcast before several times that, you know, I have two teenagers. I have a 13 year old and a 14 year old in seventh and eighth grade respectively. And this is a topic of conversation that has been pretty, pretty frequent in our home, but we didn’t start it soon enough.
I’m going to read an email that or an excerpt from an email that I received back in the fall. That email said, So the document you attached to your story is still not the original document. I need you to provide the document that you typed the story on. If you did write the original, you should have access to it and be able to attach it to the assignment.
So I’ll give you just a second to process and think about what the context for this email might have been. And you are probably guessing correctly that this is an email between my wife, myself, and my son from one of my son’s teachers. So he, we think, used AI to help write or at least edit a story and then
we got into a little bit of confusion about whether he was trying to pass that off as his own work and whether he was acknowledging use of AI or not. And that’s where this email came from. And so for someone who does what I do, being a child psychologist and certainly co-founder of an AI software company, this is essentially like being a cop and having my kid get arrested.
So to be a so-called expert in this area and have my son get flagged for potentially using AI kind of irresponsibly, this was humbling to say the least. But to me, this was not about, the point was not like what he did. I think kids are gonna do what kids are gonna do, right? In many regards. The more important thing for me was that nobody had ever taught him what responsible AI use actually looks like.
we just never had that conversation. So that was, that was on me and on my wife. And I think that’s true in a lot of households right now and in a lot of schools. And so that’s what, you know, I’m going to focus on here for the next few minutes. I want to start with some statistics though, like just a little bit of data to set the stage. So at this point we think 70%, so you know, seven out of 10 teenagers are already using
AI in some form or fashion. So it could be chat, TBT or Claude or Gemini or whatever they found. There are a lot of options out there. And most of them are doing it without any real guidance, either from the schools or from home. I think school districts are doing their best, of course, but like many cases with technology, this is one of those those times when the real world is.
moving too quickly for the policies to catch up. mean, at least to this point, I would imagine that’s going to continue to happen. Just, you know, the evolution of AI since its inception or since it really burst on the scene, you know, three or four years ago has been rapid and policy is really working to catch up. And districts and classrooms have different policies and a lot of them are pretty, pretty vague or at least variable. You know, they’re not consistent. So
As you all probably know, I’m an AI optimist, so I’m not going to harp on this idea that AI is bad or that kids shouldn’t use it. If for no other reason, then I think that ship has sailed, right? What I do think though is that we’re in a gap right now where the tools are everywhere, kids are using them, and we just haven’t done the work to teach them how to use them well. And so I’m talking about this through
I suppose a clinical lens or an education lens or parenting lens. But there’s a lot to take away here in terms of how we talk to kids during our evaluations. And part of that is incorporating questions about AI use into our screening and background forms and clinical interviews with parents and with the kids. You know, it’s like reminds me of screen time. Like we had to add in questions about screen time several years ago.
And so this is just an evolution of that process where it becomes more and more clinically relevant for the kids that we work with. So like I said, I mean, we need to be having these conversations with kids, but I don’t know that we are doing a great job of it so far. And I certainly include myself in that group, as you just heard, you know, from my story a bit.
So I think when most parents and educators hear AI and kids or AI and students, they kind of immediately jump to cheating, which is a concern. But I dug into the research a little bit, and at least thus far in anonymous surveys or confidential surveys of school-age kids, like where kids hopefully feel safe being honest.
Self-reported cheating has not gone up in any meaningful way since AI has come on the scene. Now, I don’t know if this is a matter of kids not reporting it or maybe not recognizing it or understanding that using AI is a form of cheating. We haven’t gotten that nuanced, but what we know is at least from what we have, self-reported cheating hasn’t gone up in any meaningful way. So the more…
important story for me is what happens to a kid’s when they are using AI passively. Okay, that’s not like a neutral process by any means. Using AI is replacing a cognitive process that kids need to do themselves. And over time, I think it has real effects on how well they can think independently. so, you know, academic integrity is important, of course, but actually I’m kind of zeroed in more on
one aspect or, you know, different aspects of kids thinking that, you know, are particularly important. And for the purpose of this discussion, that I’m going to focus primarily on critical thinking, critical thinking, analytical thinking, reasoning, those kinds of skills. So if you have teenagers, of course, you’re aware that critical thinking is hit or miss during the teenage years, right?
There is some neurology behind that, of course. Many of us know, you know, the prefrontal cortex is the last part of the brain to fully develop. It takes a long time. And that happens to be the part of the brain that also, you know, plays a big role in critical thinking, analytical skills, that sort of thing. The trouble is that adolescence is kind of a critical period for building these skills, like planning and organizing and critical thinking and analytical skills and reasoning.
you develop these capacities by actually using them. Kind of like, it’s like lifting weights, it’s like building muscle. You have to use those skills to develop them. And so if a kid is outsourcing a meaningful part of that process to AI during the years when the brain is most hungry for that exercise, I think that’s a pretty significant developmental concern separate from anything about
academic integrity. So by the way, you know, this is a real concern for adults as well. We already have research coming out showing that more folks that, you know, the more that folks use AI without a deliberate strategy, so kind of passively using AI as an answer generator, the worse their critical thinking becomes. So what do we do and how do we actually talk to kids about this and their AI use? Well,
I’ve started calling this the AI talk, you know, parallel to the sex talk, but nerdier and sometimes more awkward, which, which is kind of unbelievable, but, it is hard to talk about. And y’all maybe have heard me make this analogy before that, you know, the AI talk and, and use of AI is, has some parallels to the sex talk because in both cases, one,
You have to talk to kids much earlier about it than you think you need to. So the research is showing that you should be talking to kids about AI use somewhere around eight nine ten years old. Even just to introduce the idea that AI is not real. It’s making things up and you can’t trust it. Just like basic information like that. So the first thing you need to talk to kids earlier than you think you should.
Second thing, the talk is going to change at different points in the kids development and academic journey, just like the sex talk. You know, it’s not like a one time thing. You’re having it like recurring multiple times depending on where they’re at in their development. The third thing is that, you know, this is what I’ve said in the podcast before that, you know, thinking that kids or students or employees or whomever are not going to use AI is like,
hanging your hopes on an abstinence only sex ed campaign. So the third thing is that, you know, most kids are already using AI. 70 % of kids are using AI or will be soon. And the choice isn’t like whether they’re going to engage with it. It’s whether they have any guidance when they do so. So when you do have the conversation, you want to, like I said, start early, you know, don’t wait for a crisis.
You want to keep it like casual. It’s not meant to be punitive by any means. You want it to be a curious process, right? Where, you you normalize curiosity, like you bring curiosity to it and you are kind of teaching the kid how to critically evaluate AI tools. You’re hopefully not passing judgment or like developing rules only or strict guidelines around it. You want to help them kind of think through it on their own.
And really pass along that core message that AI is a tool and like any tool, it can make you better at things or it can let you avoid things and the differences in how you use it. One thing to be super aware of when you’re talking with kids is just the privacy component. So we are very tuned into this given what we do, but it’s important to emphasize with kids that privacy really matters with AI, like the things that they put into
Public AI models certainly are not private. So, you know, it’s not a therapist. It’s not HIPAA compliant or anything like that. So just making sure that they know that whatever they put in, it’s going to be used to train a model and in some form or fashion.
And then one last component of, know, what can you, what do you pay attention to? What do you watch for to trigger alarm, right? Maybe not alarm, but concern. You want to look for passive use. We’re going to talk about passive use in just a second. You want to look for, you know, accepting the output without questioning it. And you want to look for secrecy. I would add that you also want to look for like any development of
anything that resembles a relationship with AI, anything that’s like humanizing it or using it as a substitute for relationships. I am not a fan of any tool that’s going to isolate our kids more than they already are and put them on screens more than they already are and substitute for real in-person human interaction.
So let’s talk a little bit more specifically about how to use AI. So there’s a simple framework. And I will put this in the show notes because this is a little bit of a, you know, there’s a visual component to it. You know, the framework for for using AI. And I think this document that I’m looking at could be really helpful for y’all as well. So in this framework, it’s kind of a, you know, there are three components. One.
use AI for the easy stuff. So I’m OK with outsourcing the easy stuff so that you can save your brain for the hard stuff. This is true for kids, just like it is true for us as psychologists. So I’m a big believer in offloading kind of like the rote tasks so that your brain can really engage with the more challenging material. The challenge, of course, is figuring out, what is a rote task for a kid in elementary school, middle school, high school? And that’s going to change depending on the kid and the subject.
You know, having that lens is a helpful place to start. OK, so the rote tasks or the simple, the easy tasks, you you can offload those tasks. That’s OK. The second component really taps into this concept of AI as a thought partner, which I love, and that is, you know, making sure that you keep your kid’s brain in the loop. So teaching them to review the output, push back on the output.
And essentially always have control over the final draft, which is a component that we have to be aware of as well as psychologists if we’re using AI tools in our practice. You have to have control over the final draft. We’re not just like signing and sending. And then the third piece of this is looking at prompt engineering. So this sounds fancy, right? But it’s really not. So prompt engineering is essentially how to give
the AI, the instructions to make it do what you want it to do. I like prompt engineering both for adults and for kids because it’s not just a technical skill. mean, I guess there is a world where some of our kids will grow up and utilize prompt engineering in their work, like as a technical skill, like an asset. But in this context, I really like prompt engineering more as a metacognitive process or as a cognitive skill that they are developing.
Because in my mind, to be able to tell the AI what you want and be specific and have it produce good output, the kid has to actually have a pretty good idea of understanding what they need, which then means that they have a pretty good understanding of the assignment and the guidelines and the requirements and the places that they may be falling short or needing help or needing support in that process. So going through that process to prompt engineer and tell the AI what you want it to do,
can be a super valuable process even just in learning. So what is the structure of a good prompt that you can talk to your kids about?
Dr. Jeremy Sharp (19:32)
Hey, everyone. I’m really excited that NovoPsych Psychometrics is sponsoring the show. NovoPsych is a platform for psychologists who care deeply about assessment and testing and want their self-report measures to be the very best. NovoPsych has an extensive library of 150 standardized instruments with strong coverage across the presentations many of us assess every day, like disability, functional impact, autism, ADHD, and a wide range of symptom measures.
You can also use it for broad personality assessments like the Big Five or go deeper when you’re looking to understand personality pathology. What makes NovoPsych different isn’t just the range of scales, it is the quality of the experience. So I really appreciate the depth of psychometric info that it provides and the clear graphs and visualizations that make results easier to interpret and communicate. If you want to try NovoPsych psychometrics, you can access a 15 day free trial via the link in the show notes, which is
novopsych.com slash testing psychologist. That’s N-O-V-O-P-S-Y-C-H dot com slash testing psychologist.
Dr. Jeremy Sharp (20:36)
One, start with context. So you want to tell the AI what you’re working on and essentially where you are in the process. The second component is telling the AI what you already know.
OK. So your current thinking where you’re at with it. The third component is telling the AI what you’re stuck on and you want to be as specific as possible here. And then the fourth component is instructing the AI on how you want it to respond. And again, it’s very helpful to be specific and clear here. All right. So a perfect bad example here, like you would be using AI just as an answer generator. And that would be something like write me an essay about the Civil War. OK.
There’s no context. There’s no instructions. There’s really no anything. It’s just like, do this thing for me. But a much better prompt that engages active thinking and keeps the kid’s brain in the loop, has them do that metacognitive prompt engineering, be something like, I’m writing a five paragraph essay on the causes of the Civil War. My three arguments are economic differences, states’ rights, and the expansion of slavery. I think my states’ rights paragraph is weak. Don’t rewrite it.
Ask me questions to help me strengthen it. OK. So you could see you have those four components context what you already know what you’re stuck on and what you want the A.I. to do.
Like I said, I’ll put this document in the show notes so that you can download it and utilize it yourself. One last component though in this realm that I really love is the kind of like using AI in a Socratic way. So you could call this like a Socratic prompt, suppose. And it’s some version of like, don’t give me the answer. Ask me questions until I can figure this out myself. I love that for kids because again, it forces them to
answer those questions, of course, and think more deeply about the material, but it also can model for them like the kinds of questions they might ask themselves down the road as they as they get better at this skill.
With all that said, that’s big picture, you know, prompting that can be helpful for, for most kids. But like I said at the beginning, I did want to pivot into a discussion about neurodevelopmental concerns and how AI can be particularly helpful for those kids. So these are the kids that most of us are working with day in and day out somewhere between one and four and one in five kids has a diagnosable neurodevelopmental concern or mental health concern that meaningfully affects how they learn.
This is you know, that’s just the diagnosed kids many of us know that there are a lot of sub-threshold kids out there who haven’t been identified or who are just under the threshold Who could also benefit from these services? Diagnostic or school services so, you know lots of the lots of kids out there in this group and for me the upside here for AI is is even higher than with Neuro typical kids and so I want to talk about like specific groups and how we might use AI to help
So in this realm, let’s start with ADHD. So a lot of you know, two of the biggest challenges with ADHD are getting started and organizing. Now we could pick any number of things, but I’m just centering on these two concepts for now. So these are a couple of very concrete AI prompts that can help with those. So first thing for getting started, you might say something like, this is your kid, of course. I’m having trouble getting started.
Without solving this for me, that’s the key. Without solving this for me, ask me questions to help me generate ideas. OK. I’m having trouble getting started without solving this for me. Help me generate ideas by asking me questions. All right. Many of you are probably thinking, oh, that could actually work for me too. And that is very true. I have used this exact prompt in several situations over the years. Now the other component, organization. All right.
Here’s a prompt that could help with organization. Again, your kid is saying this. I’m going to list all of my ideas for this assignment. When I’m done, help me organize them into themes. So this is essentially like a structured brain dump. You know, I love this exercise in my work and I think it can benefit kids too. So they don’t have to organize anything on their own. They just have to like write down anything that might be on their mind. They can even do like speech to text. I mean, there are many ways to do it. And then
Yes, the AI, you know, when I’m done, help me organize these ideas into themes. OK, so just a couple of concrete things that you could use to help kids with ADHD. All right. Next, I want to talk about learning issues. So AI can be super helpful for learning issues. And the one overarching theme with kids with learning issues is that AI can be like an infinitely patient, personalized tutor. So.
You know, most people have finite patience. Many parents, I know, have finite patience when you’re working with your kids on homework. And many schools are overstaffed or understaffed and overworked. And so, you know, there are gaps, of course, despite best efforts. So one of those persistent challenges is kind of access to support that’s actually calibrated to where the kid is at. And it’s very hard to do one-on-one support for all the kids who need it. So AI can change that, right? So it’s infinitely patient.
It can adapt to where the kid is at. It can be super flexible. It can ask the kid the same question five different times, five different ways. On the flip side, the kid can ask the same question. And the AI does not run out of patience. It just keeps answering. There’s no like, we already covered this. Why haven’t you learned this? With specific learning disorders in particular,
Actually, one more big picture concept. know, AI can explain things in many, different ways. So I think about this with math. You know, if your kid is struggling with the method that is being taught or with a concept, AI can explain it different ways, can find YouTube videos that can help, can generate graphics or visual representations of concepts. So it can be super helpful just for flexibility and explanation.
With reading or dyslexia, it can certainly simplify passages, bring it down to a lower reading level. It can read text aloud. It can re-explain things using analogies that might fit or are easier to understand. So these are just a few ways that AI can be helpful with learning disorders as well. It can also help with writing, as we’ve talked about in getting started.
helping kids like organize their ideas into thematic groups and things like that.
So the next thing I’m going to chat about is autism. So for kids on the spectrum, two areas that AI can be super helpful are in cognitive flexibility and perspective taking. I’m just picking two, right? Both of these can be really challenging for kids on the spectrum, of course. And so in the flexibility realm, you might consider a prompt like, here are all my thoughts on this topic. Help me see this from the opposite perspective. All right. So this is one of those places kids can get stuck.
this is an easy way to break them out of that. Help me see this from the opposite perspective.
Another area that it can be helpful for kids is in perspective taking. So, you know, AI can help students kind of think through why characters or people might act in a certain way, can kind of assist in that theory of mind component, social understanding, that kind of thing. So, yeah, it meets them where they’re at. They can offer their interpretation, like, hey, this is what I think is going on in this situation, or I don’t have any idea what’s going on in the situation. Ask me questions to help me.
Help me get it. And then they can ask what they might be missing. So it’s not asking them to do something they can’t do. It’s really scaffolding a skill that can be genuinely tough for some kids on the spectrum and other kids, of course. So again, none of this is replacing targeted support, like therapy or accommodations or anything like that. But as a practice tool and an on-demand tool, I think it can be incredibly helpful.
The last thing I’m going to talk about in terms of specific prompts with groups is actually a little bit more broadly applicable. This is not just for kids with a diagnosis, but it’s for any kid or teenager who has run into difficulty with friends, like been in a fight with a friend, working through conflict, like completely unable to see outside their own perspective, which honestly is most teenagers, I think. And this can be super helpful for kids who are like too activated or embarrassed to process something with a parent or with their friends in the moment.
So it just gives a low stakes place to talk through some of these things. this is essentially like kids who are just having like, you know, conflict, like social emotional challenges. And the prompt is, I’m going to explain a conflict with my friend and how I handled it. Help me see outside my perspective. Why are we still fighting? What am I missing? Something like that. Now we can do this as well as adults, of course, but for kids, I it can be super helpful.
Now I want to emphasize, of course, like this is not therapy. The number one use of AI right now amongst the population is for therapy. I’m sure it’s helpful in some regards, but I’m not trying to replace therapy or anything like that. It’s not a replacement, like I said earlier, for relationships in the kids’ lives. But as a first step toward reflection, I think it is something. And that can be valuable for kids. So I’m going to wrap this up.
I think we’re in a genuinely important moment, not in a scary way necessarily, but in the way that any major shift in how kids learn and behave is important. The tool is here. It’s not going away. The question is, what are we going to do about it and how are we going to support our kids through it? I don’t think the answer is banning it. I don’t think the answer is handing it out to kids and hoping for the best.
I think that the answer is exactly what we do with anything else that matters, you know, as parents and educators, we pay attention to it. We talk about it. We teach our kids to use it with some judgment and we try to stay non-judgmental and curious on our parts to keep the conversation lane open. So that’s, that’s kind of my whole pitch today. I suppose is that, you know, we have real influence with our kids and the hope is that you may be walking away.
little bit more empowered about how to talk with kids, your own kids or your clinical kids about AI and have a good idea of how to support them in this area.
Click here to listen to the podcast instead.