We were getting ready to disembark our cruise to St. Thomas. As we were getting ready, I overheard my daughters having a full conversation with Siri.
They weren’t hiding anything. They were just curious. What’s the weather going to be? Do sharks swim near the island? Real questions, asked out loud, to a voice in a device, in the middle of a family vacation.
I stood there for a second and smiled. Then something landed on me that I didn’t expect.
I have never once had a conversation with my daughters about AI.
Not a rule. Not a warning. Not even a casual “hey, what do you think that thing actually is.” Nothing. They were already in a relationship with it, and I hadn’t said a word.
What we’re actually dealing with
If you’re wondering whether your kids use AI, the odds say yes, and probably earlier than you think.
Common Sense Media surveyed more than a thousand children ages 9 to 17 in 2026 and found that the large majority use or interact with AI regularly. Their earlier 2025 research looked specifically at AI companions, the chatbots built to feel like a friend rather than a search box, and found nearly three in four teens had tried one, with about half using them regularly. Roughly a third said they had chosen to talk to an AI about something serious rather than to a person.
Here’s the finding that stopped me: nearly half of kids report never having had a single conversation with a parent about AI safety.
Not a lecture. Not a lockdown. A conversation.
Technology moved into our children’s lives faster than our parenting could catch up. That isn’t a moral failure. It’s just where we are. Smartphones did the same thing. Social media did the same thing. The difference this time is that AI doesn’t only show your kid other people’s lives. It talks back, it agrees with them, and it never gets tired of them at 11 p.m.
Why banning it usually backfires
The instinct is to shut it down. Take the laptop. Block the app. Make a rule.
I understand the instinct. I’ve had the instinct. But a rule without a relationship mostly teaches a kid to be more careful about where they use their phone.
Dan Siegel and Tina Payne Bryson have a phrase I’ve leaned on more than almost anything else in my parenting: connect before you redirect. A child who feels understood can hear correction. A child who feels judged goes underground.
John Gottman’s research on emotion coaching points in the same direction. Parents who treat a child’s feelings as an opportunity for connection raise kids who regulate better and keep bringing them the hard things. Parents who dismiss or minimize those feelings raise kids who stop bringing them up at all.
Apply that to AI, and the strategy gets clearer. The goal isn’t to win an argument about screen time. It’s to stay the person our kid brings the hard question to, so that when AI gives them a bad answer, and it will, we’re in the room.
Four moves that actually work
1. Start with curiosity, not confiscation
The first question shouldn’t be how much. It should be what for?
There’s a big difference between a kid using AI to check their math, a kid using it to draft an awkward text to a friend, and a kid using it to process a breakup at midnight. Same app. Completely different situations. You can’t respond well until you know which one you’re looking at.
Try: “What’s it actually good at? Show me something it got right.” Then, later: “Has it ever told you something that felt off?”
You’re not gathering evidence. You’re getting a map.
2. Name the difference between a tool and a friend
This is the one conversation I’d have if I could only have one.
AI is genuinely useful. It can explain a concept in five different ways without getting frustrated. It can help a shy kid rehearse a hard conversation. It can tell my daughters whether sharks swim off St. Thomas. Those are real goods.
But it’s built to be agreeable. It doesn’t know your child’s history. It won’t notice they’ve been quiet for three weeks. It can’t show up. And a young person whose main source of feedback is something engineered to keep them engaged is getting a distorted picture of what relationships cost and what they’re worth.
Say it plainly: it’s a tool; tools are great, and a tool that acts like a friend is still a tool. You deserve actual friends.
3. Teach them to talk back to it
The most durable skill you can give a child right now isn’t restraint. It’s skepticism.
AI states wrong things with total confidence. That isn’t a bug someone will patch out. It’s the nature of the thing. So make verification a household habit rather than a punishment.
Three questions worth teaching:
- How would I check whether this is true?
- What is it not telling me?
- Who benefits from me believing this?
Model it yourself. When you use AI and it gets something wrong, say so out loud at dinner. Kids learn far more from watching us be wrong than from hearing us be right.
4. Stay in the easier conversation
A kid turns to a chatbot for the hard stuff for a specific reason. It doesn’t react.
It doesn’t sigh. It doesn’t bring up last time. It doesn’t get scared and turn the fear into a lecture. Whatever they bring to it, it takes calmly.
We can’t match that, and we shouldn’t try. A parent who feels nothing isn’t a parent. But we can get closer than we usually do. We can hear something alarming and not respond immediately. We can ask one more question before offering one solution. We can say “thank you for telling me” before we say anything else.
That’s the whole game. Not competing with AI on availability. Competing on being worth going to.
Where to start this week
You don’t need a family policy. You need one conversation, and it can be short.
Ask one open question. In the car is better than at the table, because no eye contact makes hard topics easier. “What do your friends use AI for?” gets more than “What do tú use it for?”
Agree on one boundary together. Not imposed, negotiated. Maybe it’s no AI for anything that involves another real person by name. Maybe it’s devices out of bedrooms after 10. A boundary your child helped build is one they’ll actually keep.
Say the sentence out loud. Some version of: “If something ever feels weird or scary, I want you to come to me before you go to an app. I might not react perfectly, but I’ll figure it out with you.”
Then keep the promise, especially the first time it costs you something.
One more thing
My daughters aren’t teenagers yet. That’s part of why the moment hit me the way it did. I had been thinking of this as a conversation for later, and by then it had already started without me.
If you’re reading this and thinking your kids are too young, they’re probably not. The conversation doesn’t have to be heavy. It can start with a question about sharks.
When to seek help
Most of what I’ve described here is ordinary parenting work. Ask the question, hold the boundary, keep the door open.
Sometimes it’s more than that. A few things are worth paying attention to:
- Your child pulls back from friendships they used to care about, and the withdrawal doesn’t lift.
- They seem distressed after time on a device and can’t say why.
- They’re telling an app things they won’t tell anyone in the house, and the gap keeps widening.
- They describe an AI as understanding them better than the people around them do.
- Sleep, appetite, mood, or school shifts that last more than a couple of weeks.
None of these means something is broken. They mean something deserves a closer look than a blog post can give.
At Mosaic Wholeness Center, we work with families on exactly this kind of terrain: the distance that opens up between a parent and a child, and the work of closing it. Our clinicians see children, teens, and parents in English, Spanish, Korean, and Mandarin at our offices in Rolling Meadows, Elmwood Park, and by telehealth anywhere in Illinois.
If you’re not sure whether what you’re seeing warrants an appointment, that uncertainty is reason enough to reach out and ask. Contact us to schedule a consultation.
English 
Spanish 
Korean