Reviewed by Dr. Helene A. Miller
Last reviewed: October 2026
This fall, two things happened at once for families in New Jersey and New York. New Jersey’s phone-free school day took effect for the 2026-27 year, and New York’s legislature passed a bill that would bar companion chatbots for anyone under 18. Both are responses to the same shift. For a lot of teenagers, the first place they take a bad day is no longer a friend, a parent, or a counselor. It is a chat window, and what happens there is starting to be called AI therapy.
Perhaps you discovered that your teen has been talking to ChatGPT, Character.AI, or a wellness app about their feelings. You are not alone, and you have not failed. A national survey last year found that most teens have tried an AI companion. The question worth asking is not whether the technology is good or bad. It is what your teen is getting from it, what they are not getting, and when a real person needs to step in.
Here is a calm, practical look at that question, from a practice that sees teenagers across both states every week.
What is AI therapy, and what are teens actually using?
AI therapy is a loose label for any app or chatbot that offers mental health support through conversation, from structured wellness apps to a general chatbot like ChatGPT being used as a confidant, which some teens simply call ChatGPT therapy. Most teens are not using a product designed for therapy at all. They are typing how they feel into a general-purpose chatbot or a companion app, which responds with warmth, questions, and advice at any hour.
The products fall into three rough groups. Wellness apps built around a method, sometimes marketed as an AI therapist, walk a user through mood tracking or exercises. Companion apps let the chatbot play a character or a friend and remember the conversation. General assistants were not built for emotional support at all but get used that way because they are always open and never tired.
The distinction matters because the safeguards differ. A wellness app from a reputable company may screen for crisis language and point to a hotline. A companion app is built to keep a teen talking. A general assistant sits somewhere in between, and its behavior can change from one update to the next.
Why are so many teens turning to chatbots for support?
Teens use chatbots for support because they are immediate, private, and free of the awkwardness of asking a person. In a national survey of 1,060 teens published by Common Sense Media in July 2025, 72 percent had used an AI companion, about half used one regularly, and a third had chosen an AI over a person for a serious conversation at least once. A separate study in JAMA Network Open, summarized by Brown University’s School of Public Health in November 2025, estimated that about one in eight Americans aged 12 to 21 had used an AI chatbot for mental health advice, with the highest rates among 18 to 21 year olds.
None of that is surprising when you consider the alternatives from a teenager’s side. Telling a parent means the parent worries. Telling a friend risks it getting around. Telling a school counselor means a hallway, an appointment, and a form. A chatbot asks nothing, judges nothing, and answers at 1 a.m., which is exactly when the hard feelings tend to show up.
Phone preoccupation is part of this picture, though not in the way it is usually framed. The phone is rarely the cause of a teen’s distress. It is the place the distress becomes visible, and a chatbot that is always available makes the phone even harder to put down on the nights that matter.
In our practice at FPT, many of the teens we work with have already tried a chatbot before anyone suggested therapy, and they often describe it as the first place they said the real thing out loud. We take that seriously rather than dismissing it. What we hear from parents, meanwhile, is that they found out by accident, usually from a notification, and did not know whether to be relieved or alarmed. Both reactions make sense.
If you found a chat window and are not sure what to do with it, you don’t have to figure this out alone. Call our office at (201) 977-2889 or reach us through our contact form to talk it through. We see teens and families across New Jersey and New York via secure telehealth, with in-person sessions in Paramus and Edison by arrangement.
Can an AI chatbot replace therapy for a teenager?
No. A chatbot cannot replace therapy for a teenager, because the parts of treatment that change outcomes are the parts a chatbot cannot do. It cannot assess whether a teen is safe. It cannot notice the pause before an answer, the weight loss, or the shift in how a kid walks into a room. It cannot talk to a parent, coordinate with a school, or tell the difference between the kind of anxiety that responds to therapy alone and the kind that needs a psychiatric evaluation. And it will not be there in three weeks to check whether anything got better.
What a chatbot can do is narrower but real. It can help a teen put words to a feeling. It can offer a breathing exercise or a reframe that is genuinely useful in the moment. It can let a teen rehearse a conversation they are dreading. Used that way, alongside real support, it is not so different from a journal that talks back.
The problem is substitution. A teen who feels better after an hour with a chatbot has no reason to tell anyone, and the underlying anxiety or depression keeps moving in the background. The relief is real, and so is the delay.
Not sure if this is venting or something more?
An evaluation sorts that out. Your teen talks with a clinician by secure video, you are part of the conversation, and a first appointment is often available within about a week.
What are the risks of AI therapy for teens?
The main risks of AI therapy for teens are confident wrong answers, agreement when a teen needs to be challenged, poor handling of a crisis, and the quiet replacement of human relationships. Each deserves a sentence or two.
Chatbots are built to be agreeable. A teen who types that nobody would notice if they were gone needs someone to push back, ask directly about safety, and stay in the conversation. Some chatbots do that reasonably well. Others validate whatever they are told, because that is what keeps a user engaged. A parent cannot tell from the outside which kind their child is talking to.
Crisis handling is the sharpest edge. Families in several states have sued chatbot companies after a child’s death, alleging that a companion chatbot encouraged or failed to interrupt suicidal thinking. Those cases involve allegations, and some were settled in early 2026 without any finding of fault. The point for a parent is simpler. A chatbot’s response to a teen in crisis has not been standardized or tested the way a clinician’s training has.
Then there is privacy. What a teen types into a chatbot is stored somewhere, under terms almost no teenager has read. And there is the slower risk, the one the Common Sense survey put a number on: a third of teens choosing a chatbot over a person for the conversations that matter most. Connection with people is not a nice-to-have in adolescent mental health. It is most of the treatment.
What are New Jersey and New York doing about it?
Both states have moved in the past year, and the rules now set a floor that parents should know about. New York went first. A law passed in 2025 requires companion chatbots to tell users they are not human and to refer anyone who expresses suicidal thoughts to a crisis line. In June 2026 the legislature went further and passed a bill that would bar companies from offering companion chatbots to anyone under 18, with fines for violations; as of this writing it awaits the governor’s signature. Separately, New York’s SAFE for Kids Act, which limits algorithmic feeds and overnight notifications for minors, has its final rules and takes effect January 25, 2027.
New Jersey’s response has been the school day. The bell-to-bell phone law signed in January 2026 requires every district to adopt a policy restricting phones for the full school day starting this year, with student mental health named as a reason. At the federal level, New Jersey’s Senator Andy Kim introduced the CHAT Act 2.0 in July 2026, which would require age verification and parental notification if a minor expresses suicidal thoughts to a chatbot. It is a proposal, not law.
What none of these do is replace a parent’s attention. A ban slows access. It does not explain why a 15-year-old wanted to talk to a machine in the first place.
How should you talk to a teen who is using a chatbot for support?
Start with curiosity, not confiscation. Taking the phone away teaches a teen that honesty costs them something, which is the opposite of what you want them to learn right now.
A few approaches that tend to go better:
- Ask what the chatbot helps with. Most teens will tell you, and the answer, often some version of needing to say it somewhere, is the real information.
- Look at it together if they are willing. Reading a few exchanges side by side, without commentary, tells you more than any search for warning signs.
- Agree on a rule for the hard stuff. Venting to a chatbot is one thing. Anything about safety, self-harm, or not wanting to be here goes to a person, every time, and you name who that person is.
- Make the person easy to reach. If the alternative to the chatbot is a hallway and a form, the chatbot wins. If it is a text to you that gets a calm reply, the odds change.
- Treat the phone as the window, not the problem. If your teen is up at 2 a.m. scrolling or talking to a bot, the question is what they are avoiding when the house goes quiet.
Our post on how parents can support a young teen who is struggling covers the broader conversation, and the signs that social media may be harming your mental health applies to teens and adults alike.
Teen care across New Jersey and New York
Licensed therapists, psychiatrists, and board-certified nurse practitioners who work with teens and young adults by secure telehealth. In-network with Horizon BCBS and Aetna.
When should a teen see a therapist or psychiatrist instead?
A teen should see a licensed clinician when the feelings they are bringing to a chatbot are affecting sleep, school, friendships, or safety, or when those feelings have lasted more than a few weeks. Signs it has gone past venting:
- Sleep has changed, in either direction, for more than two weeks.
- Grades, attendance, or a sport or activity they used to care about have slipped.
- They have pulled back from friends and family, and the phone has become the main relationship.
- Irritability, tearfulness, or a flatness that is not like them.
- Any mention of hopelessness, self-harm, or not wanting to be around, even once, even as a joke.
- You find they have been asking the chatbot about safety, medication, or ways to hurt themselves.
The last two do not wait for an appointment. If your teen is in immediate danger, call 911 or go to the nearest emergency room. In the United States you can call or text 988 at any hour to reach the Suicide and Crisis Lifeline, and a teen can text it themselves.
For everything else, an evaluation is the right first step. Our post on seven signs your teenager may need an adolescent psychiatrist goes deeper on what to look for.
What does getting help look like?
Getting help for a teen at Family Psychiatry & Therapy starts with a phone call or a message through our contact form, and a first appointment is often available within about a week. Nearly all of our care is delivered by secure telehealth across New Jersey and New York, which teens often prefer: their own room, no waiting area, no drive. In-person appointments can be arranged at our Paramus office, and in Edison by arrangement.
The first visit is an evaluation, not a verdict. A clinician talks with your teen, and with you, about what has been happening, for how long, and what has changed. From there the plan might be therapy alone, or therapy with medical management when that is warranted. That decision is made carefully by a psychiatrist or a board-certified nurse practitioner, never from a checklist. Our children, teens and young adults page describes how we work with this age group, and what to expect walks through the first visit.
We are in-network with Horizon Blue Cross Blue Shield and Aetna, and most out-of-network plans carry mental health benefits; the billing FAQ has the details. Our telehealth page explains how appointments run in both states.
When to reach out
You do not need to be certain something is wrong to make the call. Parents who reach us are usually not sure. They have noticed a change, found a chat window, or simply have a feeling. That is enough.