Kids Use AI a Lot. They Don't Love It.

Originally written in Traditional Chinese: 孩子用得最多的 AI，不一定是他喜歡的

The AI tools kids use most may not be the ones they actually like.

On October 27, Common Sense Media released its 2026 census of 1,211 U.S. kids aged 9 to 17. The share who have used an AI chatbot or companion was 36% at age 9, 55% at 11 and 73% at 13, even though the most popular chatbots officially require users to be at least 13. As the report puts it, "Age gating doesn't appear to be working among the general population."

One number stopped me. Among teens who use AI, only 11% say they enjoy it "a lot". For tweens it is 15%. Of every type of screen media the survey asked about, AI shows the widest gap between use and enjoyment. Used a lot. Not loved.

The same day, The Jed Foundation published a survey of more than 5,500 middle and high school students. 58% use AI for schoolwork. 57% use it when they feel sad, stressed, lonely or overwhelmed. Yet, as The National News Desk reported, just 8% said they specifically sought help with their feelings from AI in those situations.

Put the two together and most people see children growing dependent on AI. I see something else. My guess is that many kids open AI not because they love it, but because it is there. It is awake at midnight. It doesn't sigh.

A child opening AI when they are upset is not necessarily a child who wants to talk to AI. It may be a child who can't think of anyone else to talk to.

What kids are missing is not a better AI. It's a person they think of first when things go wrong.

To be clear, this is not a claim that AI use harms children. JED notes its data are not nationally representative and cannot show whether AI use causes distress or distressed students seek out AI. Its author, Kayla Anderson, JED's vice president of research and evaluation, says AI may offer young people an immediate and accessible outlet, but "it cannot replicate the developmental, protective, and necessary benefits of human connection."

Try this with your child this week. Imagine a good friend who just failed a test. You and your child each write three lines of comfort. Then ask an AI to write some too. Put them side by side and ask: "If you were that friend, which one would you want to receive, and why?" The AI's version is often the most polished. But it doesn't know what happened to your friend last week, or whether they want comfort or a laugh. Only people know that. Then ask one more question: "When you're upset, who is the first person you want to find?" Just listen.

Teachers can run the same exercise in class. Professionals can skip the AI-drafted message the next time a colleague is struggling and type three imperfect lines instead.

Get the Human-First practice card

Four steps, ten minutes, with copy-ready AI prompts. The card and my weekly letter are currently written in Traditional Chinese.

👉 Subscribe free to 碩學苑 (Traditional Chinese)

Follow in English

I also post in English on LinkedIn.

👉 Follow Haishuo Lee on LinkedIn

References