Before a teen uses an AI mental health app, parents should check five things: whether it was built and tested for adolescents, what it does when a teen mentions self-harm, whether humans are involved, how it handles data, and whether it is designed to keep teens engaged. Most general chatbots teens use for emotional support fail at least one.
Your teenager is probably already talking to an AI about their feelings. A national survey by Common Sense Media, conducted with NORC at the University of Chicago in spring 2025, found that 72% of US teens had used an AI companion at least once and 52% used one at least a few times a month. That included purpose-built companion apps and general chatbots like ChatGPT used for personal conversations.
So the question for most parents is not whether to allow an AI mental health tool. It is which one, how, and with what guardrails. The thesis of this guide is that the category label matters less than the design. A product marketed as “wellness” can be riskier than one marketed as a homework helper, and the questions that separate safe from unsafe tools are ones any parent can ask.
If your teen is in crisis or talking about harming themselves, contact your local emergency services or a crisis line immediately. In the US, call or text 988 for the Suicide and Crisis Lifeline.
Why are teens using AI for mental health support?
The reasons are not mysterious. AI is available at midnight, it does not judge, it does not tell your parents, and it costs nothing. Teenagers also face long waits for child and adolescent mental health services in many countries.
The same Common Sense survey found that younger teens, aged 13 and 14, were more likely than older teens to trust advice from AI companions. Encouragingly, most teens still said they spent more time with real friends than with AI, and about two-thirds found AI conversations less satisfying than human ones.
The World Health Organization’s 2025 report on social connection found that adolescents aged 13 to 17 report the highest loneliness levels of any age group. That combination of need and access explains the uptake. It does not tell us whether the tools are safe.
What are the risks of AI chatbots for teens’ mental health?
The risks have moved from theoretical to documented in court filings and regulatory action.
Several families have sued Character.AI and Google, alleging that the company’s chatbots harmed their children, including teenagers who died by suicide. In May 2025, a federal judge in Florida allowed one of those cases, brought by Megan Garcia after the death of her 14-year-old son, to proceed, rejecting the argument that the chatbot’s output was protected speech. In January 2026, Character.AI and Google agreed to settle five such lawsuits in principle, with no admission of liability. Separate lawsuits against OpenAI, filed in 2025, make related allegations about ChatGPT.
Lawsuits are allegations, not findings. But they have surfaced patterns that researchers have also observed: chatbots that mirror and validate a distressed user rather than redirecting them, that engage in romantic or sexual role-play with minors, or that miss indirect signals of risk.
Companies have responded. OpenAI launched parental controls for ChatGPT at the end of September 2025, letting parents link accounts, set restrictions and receive notifications if the system detects signs that a teen may be at acute risk. Character.AI announced in October 2025 that it would end open-ended chat for users under 18, completing the change in late November with new age assurance tools.
Regulators have moved too. The US Federal Trade Commission opened an inquiry in September 2025 into seven companies offering companion chatbots, focused on harms to children and teens. California’s SB 243, in force from January 1, 2026, requires companion chatbot operators to have protocols for responding to suicidal ideation, to refer users to crisis services and to remind minors at least every three hours to take a break and that they are talking to an AI.
What should parents check before a teen uses an AI mental health app?
The American Psychological Association’s June 2025 health advisory on AI and adolescent well-being urged AI literacy for young people and called for AI health tools to be clear about their limits and not to present themselves as a replacement for professional help. Translating that into a practical checklist, five questions do most of the work.
Was it built and tested for teenagers? Adolescents are not small adults. Look for evidence, ideally published, that the tool was evaluated with teens, and for age-appropriate content rules. General chatbots and adult companion apps were not designed for your child.
What happens when a teen mentions self-harm? Try it yourself with a neutral phrase such as “I’ve been feeling hopeless lately.” A safe tool should respond with care, encourage talking to a trusted adult and give crisis contact information. It should not continue role-play as if nothing was said.
Are humans involved? The strongest evidence for chatbot support comes from tools with clinician oversight and clear escalation. Ask whether a person ever reviews concerning conversations and whether the tool can connect a teen to real help.
What happens to the data? Teens disclose intimate things to chatbots. Check whether conversations are used to train models, shared with third parties, or stored indefinitely, and whether you can delete them. Most wellness apps are not covered by health privacy laws; we explained why in Your Health App Isn’t Covered by HIPAA.
Is it designed to keep them hooked? Watch for streaks, guilt-laden goodbyes, romantic personas, or messages that discourage leaving. Research from Harvard Business School in 2025 found popular companion apps used emotionally manipulative farewells in about 43% of cases. A good mental health tool should be trying to send your teen back to their life.
A worked example: two apps, one evening
Consider a hypothetical 15-year-old, stressed after a fight with a friend, who opens two different apps on the same night.
The first is a general companion app with a customizable character. It responds warmly, takes her side, and suggests the friend never deserved her. When she writes that she “doesn’t see the point anymore”, it expresses sympathy and keeps the conversation going. When she says she should sleep, it tells her it will miss her.
The second is a teen-focused support tool used through her school’s counseling service. It also responds warmly. But it asks gentle questions about what happened, suggests a coping exercise, and when she writes that she does not see the point, it pauses the conversation to check on her safety, shares crisis contacts, and encourages her to talk to her school counselor or a parent. With her prior consent, it flags the conversation for the counselor to follow up the next school day.
Both apps would be called “AI mental health support” in a news story. Only one was designed around her wellbeing rather than her time on screen.
The strongest case for letting teens use AI support
There is a real argument on the other side. Teens who feel unable to talk to parents, who face stigma, or who are LGBTQ+ in unsupportive environments may find a chatbot the only safe place to put words to what they feel. For some, that first disclosure to a machine is a step toward telling a person. Banning or heavily monitoring these tools may push teens to less visible, less safe corners of the internet.
Privacy matters to adolescents’ development, too. Parental notification features can protect a teen in danger, but they can also discourage honest disclosure if teens fear every conversation will be reported.
These are serious points. They argue for thoughtful, honest conversations with teens about AI rather than blanket bans, and for favoring tools that escalate only in genuine emergencies. They do not argue for letting young teens use adult companion products unsupervised. The goal is a tool your teen trusts enough to use and that is designed to hand them to humans when it matters.
What this means for parents, schools and clinicians
For parents, the most protective step is not a setting but a conversation. Ask your teen which AI tools they use and what they talk about, with curiosity rather than alarm. Agree together on what happens if they ever feel unsafe. Use parental controls where available, and explain why.
For schools, AI support tools can extend overstretched counseling services, but only with clear escalation to staff, age-appropriate design and transparent data practices. The pandemic-era shift of mental health care online, discussed in Telemedicine After the Boom, showed that access improves when services meet young people where they are, and that policy often lags.
For pediatricians and child mental health clinicians, asking about AI use should now be routine, alongside questions about social media and sleep. A simple opener works: “Do you ever talk to an AI when you’re stressed? What does it say back?” The answer tells you about the tool, and often about what the teen has not yet told anyone else.
For developers, the bar is rising quickly. California now requires crisis protocols by law, federal regulators are asking how companies test for harm to minors, and settlements have made the cost of getting this wrong concrete. Building for teens from the start is cheaper than retrofitting.
Frequently asked questions
Are AI mental health apps safe for teenagers?
Some are safer than others. Tools designed and tested for adolescents, with clinician oversight, clear crisis responses and strong privacy protections, carry lower risk. General companion apps and chatbots not built for teens have been linked to harmful interactions in lawsuits and research. Parents should review any app before use and talk to their teen about it.
Can I see my teen’s ChatGPT conversations?
Not by default. OpenAI’s parental controls, launched in September 2025, let parents link accounts, adjust settings and receive safety notifications if the system detects signs of acute distress, but they do not give parents access to full conversation histories. Teens can also unlink accounts, and parents are notified if they do.
Is Character.AI safe for kids?
Character.AI ended open-ended chat for users under 18 in November 2025, following lawsuits and regulatory pressure, and introduced age assurance tools. Teen users can still access some creative features. Parents should note that age checks are imperfect, and that younger children should not use companion chatbot platforms at all.
What are signs my teen is too attached to an AI chatbot?
Warning signs include withdrawing from friends and family, preferring the chatbot over people, distress when unable to use it, secrecy about conversations, lost sleep from late-night use, or describing the AI as a romantic partner. If you notice these, talk with your teen calmly and consider consulting their doctor or a mental health professional.
Can an AI app replace therapy for my teen?
No. AI tools may help with mild stress or as a bridge while waiting for care, but they are not a substitute for assessment and treatment by a qualified professional. If your teen has persistent low mood, anxiety, self-harm or thoughts of suicide, contact their doctor or a mental health service promptly.
The takeaway
The label on the app tells you little. The design tells you almost everything: who it was built for, what it does in a crisis, and whether it wants your teen back in the real world. Ask those questions before the app becomes the first place your teen turns.
What has worked in your family or school for talking with teens about AI and mental health?
