A teen psychologist shares the #1 warning sign your child is emotionally dependent on AI

I worry less about how many minutes a teenager spends using AI and more about the role AI is beginning to play in their life. A major warning sign is when AI starts replacing people.

Parents should pay attention if their teen would rather talk to a chatbot than a friend, parent, or therapist. If they become distressed when they cannot access it or repeatedly turns to AI for reassurance or validation. It’s a problem if your child becomes more socially withdrawn, keeps their AI interactions secret, or describes the chatbot as the only one who really understands them.

Adolescence is when we learn how to tolerate disagreement, rejection, and imperfect relationships. An AI chatbot can be endlessly available, patient, and validating. Real people aren't. If a teenager begins choosing the predictability of AI over the complexity of human relationships, that's something parents should take seriously.

My first recommendation would be: don't start by taking the phone away. That generally doesn’t go well. Start by getting curious.

Ask your teenager what they talk to the chatbot about and, more importantly, what they are getting from that relationship. Is it reassurance? Someone available at 2:00 in the morning? A place to talk without feeling judged? The answer tells you something important about what may be missing in their offline life.

Then I would be very clear about one boundary: AI can be a tool, but it cannot be your therapist. A chatbot doesn't truly know your child, cannot observe changes in behavior or affect, cannot reliably assess risk, and has no clinical responsibility for what happens next.

If a teenager is regularly discussing depression, anxiety, or other topics that indicate there is a problem with AI, I would involve a mental health professional. And if there is any concern about immediate safety, parents need to treat that as a real mental health issue—not simply a technology problem.

The goal isn't just to remove the chatbot. It's to understand why they needed it so much in the first place.

I think that is a very real concern, although this precedes AI.

What worries me is that we're removing friction from childhood and adolescence. A teenager feels uncomfortable, bored, rejected, or anxious, and within seconds they can ask a device, “Why do I feel this way?” “What should I do?”

Boredom can lead to creativity. Uncertainty teaches us how to make decisions. Conflict teaches us how to repair relationships. Sitting with an uncomfortable emotion teaches us that feelings rise and fall and that we can tolerate them.

I don't want teenagers growing up believing that every uncomfortable feeling needs to be immediately explained, solved, or eliminated. Learning to tolerate all feelings is the goal.

Healthy use might look like asking AI to explain a difficult math concept in another way, brainstorming ideas for a science project, generating questions for a study guide, or getting feedback on an essay the student has already written. The teenager is still thinking, creating, and making decisions.

Unhealthy use is when AI starts doing the developmental work for them: writing the entire assignment, deciding what they believe, constantly reassuring them, telling them whether their friends like them, or becoming the first place they go whenever they're sad, lonely, or angry.

I like to ask 3 questions: “What did you think before you asked AI?” “Do you agree with the answer it gave you?” and “How do you know the answer is accurate?”

I don't think eliminating AI from schools altogether is realistic or desirable. These students are going to enter universities and workplaces where AI will be part of everyday life. We need to teach them how to use it intelligently rather than pretend it doesn't exist.

Students still need to learn how to write an essay, solve a problem, formulate an argument, and develop an original idea without immediately outsourcing those processes. You wouldn't allow a calculator while teaching a child basic arithmetic and then conclude that calculators should never be used again. The developmental stage matters.

Yes. I support reasonable limits on phones in schools because constant notifications and social media can interfere with attention, learning, and face-to-face interaction. But removing the phone only removes the coping mechanism. It doesn't necessarily address the reason the student needed the coping mechanism.

If a teenager has been disappearing into their phone because they're lonely, socially anxious, being bullied, or uncomfortable interacting with peers, taking away the phone may suddenly expose all that distress.

Technology can sometimes be the problem, but sometimes technology is the place a teenager has been hiding from another problem.

I would have device-free meals and regular device-free family time. Phones would charge outside the bedroom overnight. I would strongly discourage teenagers from using AI for therapy, companionship, or emotional relationships. I would teach them never to assume an AI answer is true simply because it sounds confident, and never to share highly personal, medical, or identifying information without understanding where that information may go.

And I would deliberately protect some boredom.

Not every car ride needs entertainment. Not every uncomfortable emotion needs an answer. Not every awkward social situation needs an AI-generated script.

Finally, parents must follow some of these rules themselves. It's difficult to tell a teenager to put the phone away while we're answering emails at dinner.