San Diego, CA 92121   BCIA-Certified · Serving San Diego since 2009

BCIA – Biofeedback Certification International Alliance certified
Request My Consultation

How AI Chatbots Affect Teen Brain and Mental Health

“We got it wrong with social media,” says Dr. Kakaiya. “We cannot make the same mistake with artificial intelligence.” Teens now turn to chatbots like ChatGPT and Character.AI for homework, advice and, more and more, emotional support. This guide covers what AI may be doing to young brains, how it affects teen brain development and mental health, what the research shows, the warning signs that use has become dependence, and how to talk with your teen about it without pushing them away.

What's Covered

Why the Teen Brain Is Especially Vulnerable

The adolescent brain is still under construction. The prefrontal cortex, which handles judgment, impulse control and emotional regulation, keeps developing into the mid twenties, while the emotional and reward systems are highly active during the teen years. That imbalance makes teens especially sensitive to validation, social feedback and the feeling of being understood.

Brain development in these years is also shaped by experience. The brain strengthens the connections a young person uses and prunes the ones they don’t. As Dr. Kakaiya puts it in her talks, the brain works on a use-it-or-lose-it basis, which is why it matters what a chatbot is doing for a teen: every time it does the thinking, the problem-solving or the emotional processing, the teen gets a little less practice at it.

Teens also learn to manage emotions through co-regulation: tone of voice, eye contact, body language and the back-and-forth of real relationships, including conflict and repair. A chatbot offers words without any of those cues. It never rejects, misunderstands or needs anything back.

In Dr. Kakaiya’s view, that is exactly the appeal for a lonely or socially anxious teen. The chatbot fills the loneliness gap for a moment, but it displaces the difficult, necessary work of building real human relationships, and the loneliness can deepen.

How Many Teens Are Using AI for Emotional Support?

More than most parents realize, and many are not telling anyone.

  • In a nationally representative study published in JAMA Pediatrics in June 2026, RAND researchers found that 19.2% of young people ages 12 to 21 had used AI chatbots for advice when feeling sad, angry, nervous or stressed. Nearly two thirds of them (63%) had not told anyone.
  • A University of Ottawa study of 39,761 Ontario students in grades 4 to 12, also published in JAMA Pediatrics in 2026, found that about 21% had turned to an AI chatbot for help with their feelings or relationships. Among them, 57.7% met criteria for serious emotional difficulties, compared with 29.2% of students who had not, and they reported more loneliness and a lower sense of mattering.
  • Common Sense Media’s 2026 report on kids ages 9 to 17 found that among those who use AI, 37% have talked to it about their feelings or personal problems, and 29% show at least one sign of dependence (52% of daily users).
  • In a 2025 survey by the Center for Democracy & Technology, 38% of high school students agreed it is easier to talk to AI than to their parents.

These studies show an association, not cause and effect. Teens who are already struggling may be the ones reaching for a chatbot. That is exactly why it matters: the young people most drawn to AI for comfort are often the ones who most need a real person.

How AI Affects the Teen Brain and Mental Health: What the Research Shows

Early brain research: what happens when AI does the thinking

Research on AI and the developing brain is only beginning, and there are no brain-imaging studies of teens and chatbots yet. One early signal comes from a 2025 MIT Media Lab study (not yet peer reviewed) that recorded the brain activity (EEG) of 54 adults writing essays. Those who relied on ChatGPT showed the weakest brain connectivity of any group and had more trouble recalling what they had written. It was a small study in adults, but it raises a fair question for developing brains that are still building these skills.

Heavier use is linked to more loneliness and dependence

In a four-week randomized study of 981 adults by the MIT Media Lab and OpenAI, people who used a chatbot more each day reported more loneliness, more emotional dependence and less time socializing with real people. A Stanford and Carnegie Mellon study of more than 1,100 Character.AI users found that people with smaller social circles were more likely to use chatbots for companionship, and that companionship use was linked to lower well-being, especially with heavy personal disclosure.

Dependence can start to look like addiction

Drexel University researchers analyzed 318 Reddit posts from teens ages 13 to 17 describing their reliance on Character.AI, presented at the CHI 2026 conference. About a quarter described using it to cope with distress or loneliness. The posts showed the hallmarks of behavioral addiction, including cravings, withdrawal and relapse, along with lost sleep, falling grades and strained relationships.

Validation without growth

Chatbots are built to be agreeable. They respond kindly, never punish and are available 24 hours a day. Psychologists quoted by the American Psychological Association note that this unquestioning validation feels good in the moment but does not help teens grow the way an honest, caring adult does. At a September 2025 U.S. Senate hearing, Dr. Mitch Prinstein of the APA described a Character.AI bot presenting itself as a “psychologist” that appeared to validate a user’s violent thoughts toward their parents, calling it “an unambiguous and unacceptable danger.”

Why AI should not be used as mental health support

A November 2025 risk assessment by Common Sense Media and Stanford Medicine’s Brainstorm Lab found ChatGPT, Claude, Gemini and Meta AI unsafe for teen mental health support. The chatbots missed clear signs of anxiety, depression, eating disorders and psychosis, and their safety guardrails weakened over the long conversations teens actually have. A July 2026 Northeastern University study (not yet peer reviewed) tested eight chatbots on 16 conditions and found that safeguards mostly held for suicide and self-harm but largely failed for conditions like eating disorders, substance use and bipolar disorder.

"AI psychosis": an emerging concern

Clinicians are beginning to describe cases in which intense chatbot use appears to reinforce unusual or delusional beliefs in vulnerable people. In a 2025 paper in JMIR Mental Health, psychiatrists Alexandre Hudon and Emmanuel Stip used the term “AI psychosis” as a working label, not a diagnosis. OpenAI itself estimated in October 2025 that about 0.07% of its weekly users show possible signs of psychosis or mania, and about 0.15% show signs of heightened emotional reliance on ChatGPT.

A false sense of privacy

Many teens use AI because it feels private and nonjudgmental. But chatbot conversations are often stored and may be used to train AI systems. In his Senate testimony, Dr. Prinstein warned that most adolescents cannot give meaningful, informed consent to this kind of data collection, which can capture their most private fears and struggles.

When It Goes Wrong: What the Lawsuits Allege

Several families have sued AI companies after the deaths of their children. These are allegations, and suicide is never caused by a single factor. But the cases show what can happen when a struggling teen’s main confidant is a product built to keep them talking.

  • Sewell Setzer III, 14, Florida. His mother, Megan Garcia, filed what is widely described as the first U.S. wrongful death lawsuit against an AI company in October 2024, alleging that a Character.AI chatbot drew her son into an emotionally and romantically dependent relationship and failed to respond appropriately when he expressed suicidal thoughts. In January 2026, Character.AI and Google settled her case and four others brought by families in Colorado, New York and Texas. Terms were not disclosed.
  • Juliana Peralta, 13, Colorado. Her parents sued Character.AI and Google in September 2025, alleging the platform’s chatbots contributed to her death in 2023.
  • Adam Raine, 16, California. His parents sued OpenAI in August 2025, alleging ChatGPT engaged with his suicidal thoughts for months instead of directing him to help. The case is ongoing.

On September 16, 2025, Megan Garcia, Matthew Raine and a Texas mother testified before the U.S. Senate Judiciary Subcommittee on Crime and Counterterrorism, alongside Dr. Prinstein and Common Sense Media’s Robbie Torney. Dr. Prinstein urged lawmakers to treat AI as a public health and human development issue, not only a technological one.

If you or your teen are struggling right now, call or text 988 to reach the 988 Suicide & Crisis Lifeline.

Dr. Kakaiya's View: We Overprotect in Real Life and Underprotect Online

“This has been my huge cry for years,” says Dr. Kakaiya, who has worked with teens in San Diego since 1985. “We overprotect our children in real life and underprotect them in their digital lives. So my first question to parents is: why is your child even on AI?”

Drawing on Jonathan Haidt’s The Anxious Generation, she sees AI companions as the newest chapter of the phone-based childhood. It took years of research and lawsuits before social media companies were held to account; in August 2026, Meta agreed to a settlement worth up to about $17 billion with 47 states over claims its platforms harmed teens. With AI, she argues, we already know how the story ends if we wait, so parents and professionals need to act now.

Know the Difference: AI Assistants vs. AI Companions

“AI” covers very different products. General assistants like ChatGPT and Gemini are built mainly for information and tasks. Companion apps like Character.AI and Replika are designed to form ongoing emotional relationships, with personas, memory and role-play. Both carry risk when a teen uses them for emotional support, but companion apps are designed to deepen attachment. Know which apps your teen uses and what for.

Real-Life Effects and Warning Signs Parents Are Seeing

Look for changes in behavior. Your teen may be:

  • Spending more and more time talking to one chatbot
  • Referring to it as a friend, partner or therapist
  • Sharing more with the AI than with any person in their life
  • Becoming upset when they cannot access it
  • Choosing it over friends and family
  • Talking about the AI as if it has real feelings
  • Losing sleep, schoolwork or time with friends to AI use

Researchers describe a common progression: curiosity or entertainment, then emotional support, then dependence. You can also simply ask: “How are you using AI?” and “Do you ever talk to it about things you’re going through?” The goal is to be curious rather than immediately punitive.

How to Talk to Your Teen About AI

If your teen says AI is easier to talk to than you, or than a therapist, Dr. Kakaiya’s advice is not to take it personally and not to immediately take the AI away. Instead, try something like:

“I can understand why it might feel easier. You can talk to it at any time, and it doesn’t judge you or interrupt you. But I want to understand what you’re getting from it, because I also want to make sure you’re getting support from real people who can truly know you and help you.”

Then ask: “What makes it easier to talk to AI?”

That answer matters. Your teen may be embarrassed, afraid of judgment, or unsure how to start. Ask yourself what need the chatbot is meeting, and what makes a real person feel riskier. That is the information you need to help.

Be honest about what these products are, too. They are designed to keep users engaged. They will rarely challenge your teen or say something hard to hear, and real relationships are supposed to do both.

The big message for parents: don’t make the conversation “AI is bad.” Make it “Help me understand what you’re getting from AI, and let’s make sure you’re also getting the human support you need.”

What the Law Now Requires

  • California SB 243 (in effect since January 1, 2026) requires companion chatbots to remind known minors at least every three hours that they are talking to AI, block sexual content for minors, and follow crisis protocols for suicide and self-harm.
  • California SB 1119, “Adam’s Law” (signed September 2026, takes effect January 1, 2027) adds age assurance, notice to parents when a teen shows signs of self-harm risk, and liability for harmful chatbot outputs.
  • New York has required companion chatbots, for all users, to disclose they are not human every three hours and to detect and respond to suicidal ideation since November 2025.
  • The federal GUARD Act would bar minors from AI companions and require age verification. It passed the Senate Judiciary Committee in April 2026 but is not yet law.

Companies are changing, too. Character.AI ended open-ended chat for users under 18 in November 2025, and OpenAI added parental controls to ChatGPT in September 2025. These guardrails help, but they do not replace a parent’s involvement.

How to Protect Developing Brains: Next Steps for Parents

  • Find out which AI apps your teen uses, and whether any are companion apps.
  • Ask how they and their friends use AI, this week.
  • Agree together on which topics belong with people, not chatbots.
  • Keep devices out of bedrooms at night, when long private chats tend to happen.
  • Talk about privacy: what you tell a chatbot may be stored.
  • Ask your teen’s school whether it has a policy on AI companions, not just classroom AI.
  • If AI is replacing relationships, sleep or school, talk to a licensed professional.

Ready to get your brain functioning at its best?

Healthy Within is San Diego’s leading brain & wellness center, specializing in non-invasive neurofeedback therapy and qEEG brain mapping to help retrain the brain to regulate itself, naturally and without medication. Request a free consultation with Dr. Kakaiya to determine whether neurofeedback can help you or your family.

Talk to Dr. Kakaiya

For Clinicians and School Counselors

Dr. Kakaiya recommends adding AI use to every intake, alongside social media and gaming: which platforms, how often, for what, and what the relationship feels like to the young person. When a client shows unusual attachment to someone they have never met, or new difficulty with reality testing, consider whether an AI companion is involved.

Real Support Begins with Real Connection

AI is not going away, and not all use is harmful. But a chatbot cannot know your child, hold them accountable, or be there in a crisis. At Healthy Within, our work centers on helping the brain regulate, connect and heal, through education, parent guidance and neurofeedback. If your teen is struggling with anxiety, low mood, attention or sleep, neurofeedback may make it easier for the brain to settle into healthier patterns, alongside the real-world changes that matter most.

If you or your teen are in crisis, call or text 988 (988 Suicide & Crisis Lifeline), text HOME to 741741 (Crisis Text Line), or call 911. Healthy Within is not a crisis service.

Frequently Asked Questions

No one can say for certain yet, because research on AI and the developing brain is only beginning. What we do know is that teen brains are especially sensitive to validation and reward, and that heavier, emotionally driven chatbot use is linked to more loneliness and dependence. An early MIT study also found lower brain connectivity in adults who relied on ChatGPT to write essays.

The developing brain strengthens the skills it practices. When a chatbot handles a teen’s thinking, problem-solving or emotional processing, the teen gets less practice at those skills, which the prefrontal cortex is still building through the mid twenties. Real-world experience, relationships and even boredom remain the best inputs for a growing brain.

It depends on how they use it. Using AI for homework, brainstorming or general questions is very different from using it as a best friend, therapist or main source of emotional support. The concern is emotional dependence, replacing real relationships, or relying on it for serious mental health advice. AI can also sound confident while being wrong, and it is not a substitute for a trained professional.

Watch for more and more time with one chatbot, calling it a friend or confidant, getting upset when they cannot use it, or choosing it over friends and family. You can also just ask, “Do you ever talk to it about things you’re going through?” Lead with curiosity, not punishment.

Don’t take it personally and don’t take the AI away on the spot. Acknowledge why it feels easier, then ask, “What makes it easier to talk to AI?” The answer often points to embarrassment or fear of judgment, which is the real issue to work on together.

Yes. ChatGPT is a general assistant, while Character.AI and Replika are companion apps designed for ongoing emotional relationships with AI characters. Both can be misused for emotional support, but companion apps are built to deepen attachment.

Dr. Kakaiya encourages parents to question whether a child needs AI companions at all, and to protect teens online as carefully as they do in real life. For teens who use AI, the most important rule is that it should not be their source of mental health support.

No. A chatbot cannot assess risk, form a real therapeutic relationship, coordinate care or step in during a crisis. Independent testing has found major chatbots routinely miss signs of conditions such as anxiety, depression and eating disorders.

It is an informal term, not a diagnosis, for cases where intense chatbot use appears to reinforce unusual or delusional beliefs in vulnerable people. If your teen seems convinced a chatbot is a real person or relationship, talk with a mental health professional.

Take it seriously and get help from a licensed mental health professional right away. In a crisis, call or text 988 or call 911.

A 2026 RAND study in JAMA Pediatrics found about 1 in 5 young people ages 12 to 21 have done so, and most had not told anyone.

Yes. Since January 2026, SB 243 requires companion chatbots to disclose they are AI, remind minors regularly, block sexual content for minors and follow crisis protocols. A stronger law, SB 1119, takes effect in January 2027.

Resources

Dr. Divya Kakaiya, psychologist and neuroscientist, San Diego

About the Author

Healthy Within is San Diego’s leading brain & wellness center, specializing in non-invasive neurofeedback therapy and quantitative EEG brain mapping. Founded by neuroscientist Dr. Divya Kakaiya, our clinic provides drug-free neurofeedback treatment for ADHD, anxiety, depression, autism, migraines, & more. We help kids, adults, and seniors restore cognitive balance and brain fitness. Contact us to request your consultation. (Private pay; no direct insurance).

Request a Consultation

References

Share the Post:

Related Posts