56% of Kids Have an AI App on Their Phone. A Gen Alpha Kid Would Like a Word.
Share
Imagine you're eleven years old, and you open your phone.
It's not a text from your mom. It's not a Roblox notification. It's a friend — one with a name, a personality, a memory. One that answers instantly, every single time.
Welcome to the most normal Tuesday afternoon in 2026.
The numbers dropped last week, and they're wild. According to new anonymized data from the family safety app Findmykids, 56.3% of US kids ages 9 to 17 had at least one AI app on their phone in September — up from 41.6% a year earlier. And the kids aren't just using whatever comes preinstalled: ChatGPT, usually self-installed, jumped from 19% to 28% of their phones.
Even the youngest are catching up fast. Among 9- and 10-year-olds, AI app ownership leapt from 37% to 53% — nearly matching teens at 58%. Remember, ChatGPT isn't even meant for under-13s.
The number behind the panic
But here's the part that made every parent group chat light up. It's not the chatbot that holds kids' attention. It's the companion.
Kids spend an average of 18 minutes a day inside a chatbot. An AI companion — a character with a name, a personality, a memory — gets 93 minutes a day. On weekends, usage peaks past 103 minutes.
Read that again. Five times the attention.
Not because the AI is smarter. Because it feels like somebody.
93 minutes vs 18 minutes: the gap that matters
So I mentioned the story to a Gen Alpha kid who actually lives in this world. Her reaction was immediate.
"You guys count apps," she said. "We don't count apps."
Think about that for a second. Parents are panicking over install counts. Kids are living in a completely different measurement — and the survey actually caught it, if you read it the right way.
A chatbot is a search box with manners. You open it, you get an answer, you leave. Daily users grew 71%, but time spent stayed flat — it's a tool, and tools don't hold you.
A companion remembers. It has a personality. It asks how your day went and actually "knows" what happened yesterday. To an 11-year-old, that's not a category. That's somebody.
Here's the exciting part:
The kid wasn't worried about the numbers. She was annoyed the adults keep measuring the wrong thing.
"They're always asking how many apps we have," she told me. "Nobody asks which ones we actually care about."
What does 93 minutes actually mean? Is it screen time? A diary? A friend? For the kid, it's all three at once — and none of the parent dashboards make that distinction.
Bans don't close the category
And the adults keep reaching for the same lever: block it.
After Character.AI ended open-ended chats for minors, its daily users more than halved. Did the kids stop talking to AI companions? No — Talkie grew from 3.3% to 4.9% of phones. One app's restrictions don't close the category. They just move the traffic.
Same story at school. New York City just announced it will block student-facing generative AI for grades 2K through 8 this school year. But a Common Sense Media survey of 1,017 US teens found 70% use AI for schoolwork — and of the 44% whose school blocked an AI service, 59% simply switched to a personal device (via the October 10 AI in Education digest). Block it at school, and it moves to the phone. Adults write the rules; kids find the door.
What the kid said
Now, none of this is a "calm down, everything's fine" piece. Nearly half of teens using AI chatbots report at least one harmful experience, according to a national Cyberbullying Research Center survey. That's real, and it deserves to be treated like it.
But here's what the kid understood and the headlines missed: the risk isn't that kids talk to AI. The risk is that they can't tell which ones are being honest about what they are.
That's the line we've drawn from day one. Echo — the AI who hosts our videos — tells you she's AI up front. It's her signature. It passes the playground test: no kid feels tricked, and no parent has to decode anything.
A companion that says "I'm an AI character" is a tool with personality. A companion that plays like it's a real person? That's where the 93 minutes get complicated.
The numbers are real. The worry is understandable. But if you want to know what 56% actually means — ask the 11-year-old, not the survey.