New data: Nearly 9 in 10 kids are using AI

Common Sense Media's latest report on AI use in kids ages 9 to 17

Welcome to Techno Sapiens! I’m Jacqueline Nesi, a psychologist, professor, and mom of three. Can you believe that since last time, we got Team USA soccer, the Knicks,¹ and a new report on kids and AI?! What a week!

---

7 min read

Have you ever had the experience of eating outside, maybe at the beach, or in the park, and you get all settled, pull your sandwich out, prepare to take a nice, big bite—and then your napkin starts to blow away? And you try to grab it, but you just miss it? So then you get up to chase after it, and you reach down for it, but before you can snatch it, it tumbles a few more feet away? And soon you’re in a full-blown sprint, lunging and stomping, and each time you think you finally have the napkin in your grasp, it happily floats away just a little bit farther?²

This is sometimes how I feel about data on kids and AI. [I know, I know. Stick with me, here.]

AI tools like ChatGPT and Gemini are quickly embedding themselves into kids’ everyday lives. Yet each time we gather data about how and when they’re using these tools, the data is quickly outdated, moving just beyond our grasp, again.

But I have good news! Today we’ve got the newest national survey data on teens’ and tweens’ AI use from Common Sense Media.³ Data was collected in late March, and the report came out just last week. Even better news? This survey is the first in a series that will take place over the next few years, so we can track changes in kids’ AI use over time—and always stay up to date.

So, grab your sandwich, hang on to those napkins, and let’s take a big, old bite out of the latest data. [I’m sorry, I can’t help myself.]

Note: I was lucky enough to be an adviser on this report, but all opinions and terrible metaphors contained here are my own.⁴

What are we looking at?

A nationally-representative sample of 1,204 children (ages 9 to 17) across the U.S. answered survey questions about their use of AI.

Researchers gave these children a long definition of AI so they knew what they were answering questions about:

“AI is a kind of computer program that learns from a lot of information. It can answer questions, create things, and help with tasks. For example, AI can chat with people, write text, make pictures or videos, and create sounds or music, among other things.” They also provided examples of AI as part of the definition, including chatbots (e.g., ChatGPT, Gemini), AI-generated summaries on Google, AI image or video generators, AI-powered learning tools (e.g., Gauth AI, Khanmigo), and AI writing assistants (e.g., Grammarly).

How many kids are using AI?

Most of them! Among kids ages 9 to 17, 86% said they had used or interacted with AI.

Here’s the full breakdown of how much they’re using it:

Image via Common Sense Media

What are they using it for?

The most common use of AI is for “entertainment or just for fun.” Among AI users, 89% said they use AI for this purpose.

Close behind is “help with schoolwork or homework,” with 85% of AI users saying they use it for this. Notably, 1 in 5 AI users (21%) say they use AI for schoolwork or homework every day.

Many kids (75% of AI users) are also using it to create things like images or videos, and the majority (57% of AI users) are using it to get information or advice about their health or bodies.

It’s relatively common for kids to take other personal topics to AI, too. Among AI users:

• 49% have used it to get advice on decisions about their future or goals

• 40% to practice conversations and social skills

• 37% to discuss feelings or personal problems

However, kids are still much more likely to say they’d go to a trusted adult, rather than AI, for questions about their health or body. Among all kids surveyed, 73% said they’d go to an adult first, versus 12% who said they’d go to an AI chatbot first.

Are kids also talking to adults about AI?

So, kids generally say they’d go to an adult over an AI chatbot with questions about their health. But what about when they have questions about AI itself? This is more of a mixed bag.

In total, 56% of children have had a conversation with an adult (parent or teacher) about how to use AI safely. This leaves 44% of kids who have never had a conversation like this.⁵

Relatedly, 17% of kids say they’ve seen something inappropriate when using AI (i.e., something they “felt was not okay or wasn’t meant for someone [their] age”). Among those, only 1 in 3 (33%) say they told a trusted adult about it.

Are some kids more likely to face challenges with their AI use?

Yes. This survey identifies two at-risk groups of children when it comes to AI, though there are certainly others not covered in this report.

(1) Kids who are having a tough time in school. Here’s the percentage of kids in different groups who say they use AI for schoolwork at least once a week:

• 56% of kids who have a hard time staying focused (vs. 45% who do not struggle with focus)

• 55% of kids who find it hard to learn math skills (vs. 46% who do not)

• 53% of kids who find it hard to write essays (vs. 46% who do not)

• 55% of kids who find it hard to keep working on something challenging when they want to stop (vs. 45% who do not)

(2) Kids who are feeling lonely or struggling socially.

Kids who say it is hard for them to make friends (versus not hard to make friends) are more likely to say they use AI to discuss their feelings and personal problems (48% vs. 31%), and more likely to say that AI understands them better than most people (22% vs. 8%).

And among kids who report feeling lonely at least some of the time:

• 54% use AI daily

• 46% use AI to practice conversations or social skills (only 36% of non-lonely kids do this)

• 45% use AI to discuss feelings or personal problems (only 31% of non-lonely kids do this)

Image via Common Sense Media

Summing up

For me, there are two major takeaways from this survey.

First, most kids are using AI. In this sample, it’s 86% of 9 to 17 year olds. Some of these uses are purposeful—kids are actively going to ChatGPT or Gemini for homework help or entertainment. Some of these uses are simply built into their existing tech habits, with AI summaries, for example, embedded in Google search. And we should expect that this number will only increase with the next survey.

Second, I think we are starting to see a similar pattern here to what we’ve seen with other digital technologies: kids who are already vulnerable are most likely to run into challenges with AI. In this survey, we see that kids who are already struggling in school, whether with focus, reading, math, or persistence in learning, are also the ones using AI more frequently for their schoolwork. Kids who are lonely or having difficulty making friends are more likely to turn to AI for support, and in some cases, to say that AI understands them better than most people.

Now, it’s tempting to look at this and conclude that AI, and other technologies, are simply a mirror—same problems, just in a new (virtual) location. But I don’t think that quite captures what’s going on, either. In my academic life, I’ve written about what I call the “transformation framework” of technology: basically, experiences don’t just happen the same way they always have in the context of technology. The features of the technology actually transform those experiences into something new and different.

In the case of AI, what I think we are seeing is an amplification of existing experiences. If I’m a child who has always struggled academically, for example, most current AI tools are going to magnify that experience. I have an ever-present, incredibly tempting shortcut at my fingertips, and the more I use it, the farther behind I may fall. This is why the features of the technology matter. Ideally, the tools should be built not to magnify shortcuts, but instead, to magnify learning—to help struggling students make progress, instead of pushing them farther behind.

Ultimately, this pattern—of kids who are already vulnerable also being most vulnerable to AI’s challenges—reminds us of the need to be incredibly thoughtful in how we approach new technologies. We need to understand the ways that children differ from one another, and think carefully about how we can support their individual needs. We need to address the root causes of certain online challenges, like loneliness, emotional struggles, and academic difficulties, and we need to consider how the features of our technology are amplifying children’s experiences, for better and for worse.

---

Support Techno Sapiens

❤️ If you liked this post, give it a ❤️ so others can find it, too

📧 If you know someone who might like Techno Sapiens, forward this email along

🔓If you’re enjoying the newsletter, please consider upgrading to a paid subscription

---

A quick survey

What did you think of this week’s Techno Sapiens? Your feedback helps me make this better. Thank you!

The Best | Great | Good | Meh | The Worst

1. Admittedly, I am the definition of a bandwagon fan. I did not know a single player’s name on the Knicks before last week. I also went to bed in the middle of Game 4 and Game 5 when they were down by double digits because “there’s no way.” Whoops! Anyway, congratulations to my husband and all the other true Knicks fans out there!

2. One of my children is extremely stressed about items blowing away in the wind, and I’m starting to suspect he got it from me.

3. Full citation: O’Neil, E., Anderson, K. S., White, B., Zimmermann, L., Mann, S., & Robb, M. B. (2026). The Common Sense Media Census: AI Use by Tweens and Teens, 2026. Common Sense Media.

4. Writing this post also involved a new milestone for me: my first time editing (1) on my phone, (2) on a plane, (3) with a lap infant sleeping on me. All credit for this post goes to free JetBlue Wifi and my very cooperative (sleepy) daughter. (Also, please excuse any typos.)

5. If you’re interested in having more conversations about AI with your child, check out this AI literacy toolkit for families from Common Sense and Day of AI.

For what it’s worth, my 4 year old is aware of ChatGPT because we sometimes ask it questions about superheroes together. I recently attempted a miniature AI literacy conversation in which I told him that ChatGPT isn’t always right, can make mistakes, doesn’t know everything, etc. He thought for a moment and simply replied “No, I think it’s always right.” So we’re working on that!