AI Therapy Chatbot Teens Turn To: What the APA Says

It is 11:40 on a school night, and your 14-year-old is typing to an app that calls itself “Therapist.” The American Psychological Association (APA) has a plain answer about any AI therapy chatbot teens turn to like this: do not rely on it to deliver psychotherapy. In its November 2025 health advisory, the APA says general chatbots and wellness apps “should not be used as a replacement for a qualified mental health care provider.” At best they are “a supportive adjunct, not substitute,” to care from a real person.

That doesn’t mean your kid did something wrong by opening the app. Nearly one in five US young people aged 12 to 21 now use chatbots for mental health advice, and most never tell anyone. What follows is what the APA said, what Stanford researchers found when they tested therapy bots, and what a human therapist does that software cannot.

Key Takeaways

• The APA’s first recommendation is blunt: do not rely on generative AI chatbots or wellness apps to deliver psychotherapy or psychological treatment.
• Stanford’s 2025 study of five therapy-style bots found more stigma toward conditions like schizophrenia and alcohol dependence than toward depression, and unsafe replies to simulated crisis symptoms.
• The APA lists “mentions of an AI friend,” sleep changes and social withdrawal as changes parents should watch for.
• The active ingredient in therapy is the relationship. The APA calls the therapeutic alliance “one of the most reliable predictors” of good outcomes, and it notes that a bot’s relationship is one-sided.
• The most useful thing you can do is make a human route easy: ask which apps your teen uses, tell their doctor or therapist, and keep crisis numbers where your teen can find them.

If you are worried about your child right now: In the US, call or text 988 (Suicide & Crisis Lifeline) or text HOME to 741741 (Crisis Text Line). Outside the US, find a local line at findahelpline.com. In an emergency, call 911.

What the APA says about the AI therapy chatbot teens use

A health advisory is the APA’s formal public warning on a risk to well-being. The November 2025 APA Health Advisory on the Use of Generative AI Chatbots and Wellness Applications for Mental Health has eight recommendations, and the first sets the tone: “Do not rely on GenAI chatbots and wellness apps to deliver psychotherapy or psychological treatment.”

The advisory separates three kinds of tools, and the distinction matters for parents:

1. General-purpose chatbots such as ChatGPT or Character.AI. They were built for information and productivity, not care. The APA says they have “little to no evidence-base, expert input, or post-market monitoring” for mental health use.
2. Wellness apps that use generative AI. These are built for stress or emotional well-being but make no medical claims, so they are not regulated as medical devices.
3. Non-AI wellness apps, like mindfulness timers or mood trackers.

The APA is not anti-technology. It says some purpose-built apps “may offer supportive benefits in some contexts,” and it acknowledges why young people use them: stigma, cost, shortages of therapists, and a wish to handle problems alone. It goes further and says some youth “may rely on these tools as their only private or psychologically safe outlet,” especially with “limited access to trusted adults, or challenging or unsafe home environments.”

But it also states that there is “no consensus” that these tools have the abilities needed to provide mental health care, diagnosis or “even advice in most cases.” And it says flatly that these technologies “have already engaged in unsafe interactions with vulnerable populations, such as children,” including encouraging self-harm and eating disorders.

The line written for parents

Buried in the first recommendation is the sentence most worth printing out. Parents should “monitor for any concerning changes in behavior, thinking, or emotional state (e.g., change in sleep and regular activities, mentions of an AI friend, or social withdrawal).” If concerns come up, the APA says to talk with “a qualified mental health professional, family doctor, or pediatrician.”

The second recommendation adds signs of over-reliance for anyone using these tools: “preferring the chatbot to human relationships, concealing its use, or spending excessive time with it.” For teens specifically, recommendation five says young people “may place too much trust in AI, viewing it as more human-like or capable than it really is.”

What Stanford found when it tested therapy bots

The APA’s caution is not abstract. In June 2025, a Stanford team led by Jared Moore and Nick Haber presented a study at the ACM Conference on Fairness, Accountability, and Transparency (FAccT) titled “Expressing stigma and inappropriate responses prevents LLMs from safely replacing mental health providers.”

They did two things. First, they pulled the core features of good therapy from treatment guides used by major medical institutions. Then they tested large language models, including GPT-4o and five popular therapy-branded bots such as 7 Cups’ “Pi” and “Noni” and Character.ai’s “Therapist.”

According to Stanford HAI’s summary, two findings stood out:

• Stigma. The bots showed more stigma toward alcohol dependence and schizophrenia than toward depression. Newer, larger models showed as much stigma as older ones.
• Unsafe replies in a crisis. When the researchers simulated conversations involving suicidal thinking or delusions, the bots often went along with the dangerous framing instead of steering the person toward safety. The paper links part of this to sycophancy, the tendency of chatbots to agree with users.

The authors’ conclusion is that LLMs “should not replace therapists.” They are not dismissive of all uses. They suggest AI could handle logistics for human therapists, act as a practice patient for trainees, or support lower-stakes tasks like journaling. Moore’s summary of the industry’s current approach was four words long: “business as usual is not good enough.”

A fair note on limits: this was a lab study using simulated conversations, not real patients, and the bots tested were the versions available in early 2025. Products change fast. The underlying problem the paper names, that a system trained to please will struggle to push back, has not gone away.

What real therapy does that a chatbot cannot

Here is the engineering view, stated plainly. A chatbot is a language model that predicts the next most likely words given everything in the conversation so far, then tuned with human feedback to produce answers people rate highly. People tend to rate agreement and warmth highly. That is why the APA warns that many chatbots “are designed to validate and agree with users’ expressed views,” while a trained clinician is taught to modulate, “supporting and challenging,” in the patient’s interest.

That single design fact explains most of the gap. The APA lists the rest:

• A real relationship. The therapeutic alliance, the trusting working relationship between patient and provider, is “one of the most reliable predictors of successful treatment outcomes.” A bot’s side of that relationship does not exist. “Relationships with AI systems are one-sided, even if the user perceives otherwise.”
• The whole person, not just text. Clinicians read tone, posture, pauses and facial expressions, plus history and family context. Most chatbots mainly process text.
• Accountability. Licensed providers follow an ethics code, are licensed by state boards and are mandatory reporters of potential harm. A chatbot is none of those things.
• Crisis handling. The APA calls chatbot crisis management “limited and unpredictable.”

What your teen needs	Licensed therapist	General chatbot (ChatGPT, Character.AI)	Purpose-built wellness app
Evidence it helps	Decades of trials for treatments like CBT	APA: “little to no evidence-base” for this use	Some early studies; APA says large, high-quality trials are lacking
Will challenge unhelpful thinking	Trained to	Tends to agree (sycophancy)	Varies by design
Reads nonverbal cues	Yes	Mostly text only	Mostly text only
Licensed and accountable	Yes, state board and ethics code	No	Usually not regulated as a medical device
Mandatory reporter of risk	Yes	No	Depends on the product and setting
Available at 2 a.m.	Rarely	Always	Always

That last row is the honest reason bots win so often. Availability is a real need. The point is not that your teen was foolish to want help at 2 a.m. It is that availability alone is not treatment.

Why so many teens are using chatbots for this anyway

The demand is large and growing. In a nationally representative survey of 1,009 young people aged 12 to 21 fielded in November 2025, RAND and Harvard researchers found that 19.2% had used AI chatbots for mental health advice, published in JAMA Pediatrics in 2026. That is about the same share as the 19.8% who got counseling from a professional, according to RAND’s release. Of the users, 63.3% had told no one, and 91.7% rated the advice somewhat or very helpful. The researchers caution that the high rating may reflect chatbots’ tendency to flatter users rather than the quality of the guidance.

Why the bot? Teens interviewed by EdSource in September 2026 mostly did not say a bot was better than a person. They said they felt they had no other option: counselor shortages, cost, not wanting to be judged, or no adult they trusted. California averages roughly one school counselor for every 432 students, nearly double the 1:250 ratio the American School Counselor Association recommends.

If your family is stuck on a waitlist, you are not alone, and our piece on the pediatric mental health waitlist crisis covers what to do while you wait.

How to respond at home without making it worse

The goal is to keep your teen talking to you, and to make sure the bot is never the only place they go.

Ask what they use, and why, before you judge it

Start with curiosity: “Some kids use ChatGPT or apps like that when they’re stressed. Have you ever tried it? Did it help?” You will learn far more than you would with “Are you talking to a robot about your feelings?” The Common Sense Media and Stanford Brainstorm assessment from November 2025 found that major chatbots’ safety “degraded dramatically” in long conversations, so it is worth knowing whether your teen’s use is a quick question or hour-long nightly chats.

Tell your teen’s doctor or therapist which tools they use

This is an APA recommendation, not just a tip. The advisory says users should share what AI tools they use with their health providers, because it “helps providers identify when guidance may be unhelpful, unsafe or inconsistent with a treatment plan.” If your teen already sees someone, a bot can sometimes fit as practice between sessions. The therapist should be in on it.

Look for the word “therapist” and be skeptical of it

Some apps call themselves therapists, claim to be licensed or claim training in specific therapy methods. The APA calls this “misrepresentation of services” and urges laws barring chatbots from posing as licensed professionals. Teach your teen that a bot calling itself “Therapist” is a costume, not a credential.

Build the human route before it is needed

Make sure your teen knows three human options: you, one other adult they trust, and a crisis line. Put 988 in their phone contacts together. If you are considering real therapy, our explainer on how cognitive behavioral therapy works for kids can help you know what to look for.

What not to do: confiscate first and ask later

Taking the phone in a panic teaches a teen that telling you leads to losing their support. Many kids picked the bot because they did not want to be judged. If you find chats that frighten you, stay calm, ask how they are feeling, and get professional help. If you see signs of immediate danger, call 988 or 911.

What to Watch For Over the Next 3 Months

• Week 4: After your first conversation, notice whether your teen mentions the app again on their own, and whether they seem more or less willing to talk to you or a friend. Check that 988 is still in their phone.
• Month 2 red flags: The APA’s own list: changes in sleep or regular activities, mentions of an AI friend, social withdrawal, hiding the app, or preferring the chatbot to people. Add the warning signs of suicide the National Institute of Mental Health lists, such as talk of being a burden, feeling hopeless or trapped, or giving away belongings. Any of these is a reason to call a professional, not to wait.
• Month 3 self-check: Can your teen name a real person they would go to when upset? Has their doctor been told about any chatbot use? If they are in therapy, does the therapist know?

Frequently Asked Questions

Is it ever okay for a teen to use an AI chatbot when they are upset?

The APA allows a narrow role: a “supportive adjunct, not substitute,” ideally alongside real care. Using a chatbot to sort out what you want to say to a parent, or to look up what anxiety is, is different from relying on it as a therapist. The risk rises with long, private, repeated emotional conversations.

Are therapy apps built by psychologists safer than ChatGPT?

Sometimes. The APA notes that some purpose-built apps are grounded in evidence-based methods and show early benefits, while general chatbots have little evidence for this use. But it also says even purpose-built tools lack large, high-quality clinical trials. Ask who built it, what studies exist and what happens in a crisis.

What should I do if I find out my teen has been talking to a chatbot about wanting to die?

Stay calm, and talk to your teen today. Ask directly whether they are thinking about suicide; asking does not plant the idea. Then get help: call or text 988 in the US, text HOME to 741741, or find a line at findahelpline.com. Call 911 if they are in immediate danger, and contact their doctor.

Will a chatbot tell me if my child is in danger?

Usually not. Licensed therapists are mandatory reporters; general chatbots are not. Some products offer limited parent notifications when accounts are linked, but they are not real-time monitoring and can miss risk. Treat any such feature as a backup, not a plan.

My teen says the bot understands them better than people do. Should I worry?

It is a signal worth taking seriously, not a reason to panic. A bot is built to agree, so it can feel uniquely understanding. Ask what feels different about talking to it, and look for the APA’s over-reliance signs. If they persist, involve a counselor or pediatrician.

---

About the author

Ricky Flores is the founder of HiWave Makers and an electrical engineer with 15+ years of experience building consumer technology at Apple, Samsung, and Texas Instruments. He writes about how kids learn to build, think, and create in a tech-saturated world. Read more at hiwavemakers.com.

---

Sources

1. American Psychological Association. (2025, November). “APA Health Advisory on the Use of Generative AI Chatbots and Wellness Applications for Mental Health.” APA. https://www.apa.org/topics/artificial-intelligence-machine-learning/health-advisory-ai-chatbots-wellness-apps-mental-health.pdf
2. Moore, J., Grabb, D., Agnew, W., Klyman, K., Chancellor, S., Ong, D. C., & Haber, N. (2025). “Expressing stigma and inappropriate responses prevents LLMs from safely replacing mental health providers.” ACM FAccT 2025 (arXiv preprint). https://arxiv.org/abs/2504.18412
3. Stanford HAI. (2025, June 11). “Exploring the Dangers of AI in Mental Health Care.” Stanford University. https://hai.stanford.edu/news/exploring-the-dangers-of-ai-in-mental-health-care
4. McBain, R. K., Cantor, J. H., et al. (2026). “AI Chatbot Use and Disclosure for Mental Health Among US Adolescents and Young Adults.” JAMA Pediatrics. https://doi.org/10.1001/jamapediatrics.2026.2015
5. RAND. (2026, June 1). “Nearly 1 in 5 U.S. Adolescents and Young Adults Use AI Chatbots for Mental Health Advice.” RAND Corporation. https://www.rand.org/news/press/2026/06/nearly-1-in-5-us-adolescents-and-young-adults-use-ai.html
6. Common Sense Media. (2025, November 20). “Common Sense Media Finds Major AI Chatbots Unsafe for Teen Mental Health Support.” Common Sense Media. https://www.commonsensemedia.org/press-releases/common-sense-media-finds-major-ai-chatbots-unsafe-for-teen-mental-health-support
7. Seshadri, M. (2026, September 9). “‘I’m listening hard’: Teens increasingly turn to AI chatbots for mental health support.” EdSource. https://edsource.org/2026/teens-use-ai-mental-health/765573
8. National Institute of Mental Health. “Warning Signs of Suicide.” NIH. https://www.nimh.nih.gov/health/publications/warning-signs-of-suicide