Signs of AI Dependency: Is Your Child Too Attached to a Chatbot?

One chat that mentions loneliness is not a crisis. Three months of a child who reaches for the bot before reaching for anyone else might be. The most reliable signs of AI dependency are patterns over time, not single red flags, and that is exactly where many parents go wrong. In a 2025 interview study, parents reviewing real youth-chatbot conversations tended to treat one event, such as a mention of suicide or flirting, as the highest risk. Developmental psychologists looked instead for “patterns over time, such as repeated references to self harm or sustained dependence.”

This guide borrows a tool clinicians have used for two decades, the six-component model of behavioral addiction, and shows what each component looks like when the object is a chatbot. It ends with a checklist you can actually run for three months.

Key Takeaways

• A 2026 Drexel study of 318 Reddit posts by self-identified 13- to 17-year-olds found all six components of behavioral addiction in teens’ descriptions of Character.AI use, along with lost sleep, falling grades and strained relationships.
• The APA lists three over-reliance signs: preferring the chatbot to people, concealing its use, and spending excessive time with it. For parents, it adds changes in sleep, “mentions of an AI friend,” and social withdrawal.
• In a four-week randomized trial with 981 adults, the people who chose to use a chatbot more had worse loneliness and dependence outcomes, whatever version they were given.
• Dependence can feel like efficiency. Young adults in a 2026 study described “feelings of atrophy in abilities” after handing tasks to AI.
• Watch the trend line. A single worrying message calls for a conversation; a rising pattern across weeks calls for a plan and, sometimes, a professional.

What “Emotional Dependence on a Chatbot” Means

Emotional dependence on a chatbot is a pattern in which a child relies on an AI system to regulate feelings, and that reliance starts to crowd out people, sleep, school or other ways of coping. It is not the same as frequent use. A kid who chats with a homework bot every night is a heavy user. A kid who feels anxious and incomplete when they cannot talk to a character is showing something different.

Researchers do not yet have a validated clinical diagnosis for this, and nobody should claim one. What they have is a well-tested framework from other behaviors. In 2005, psychologist Mark Griffiths proposed in the Journal of Substance Use that any behavioral addiction, from gambling to gaming, can be described with six components: salience, mood modification, tolerance, withdrawal, conflict and relapse. The model is widely used in research on problematic gaming and social media.

A team at Drexel University, led by Afsaneh Razi with first author Matt Namvarpour, applied that lens to chatbots. Their paper, presented at the ACM CHI conference in April 2026, analyzed 318 Reddit posts by users who said they were 13 to 17 and wrote about their dependence on Character.AI. About a quarter of the posts described using it for emotional or psychological support, such as coping with loneliness or distress. All six components appeared.

Two limits matter. Ages were self-reported on Reddit, and the sample consisted of teens who chose to post about dependence, so it shows what dependence looks like, not how common it is. And it covers one app.

The Six Signs of AI Dependency, Translated for Parents

Here is the model applied to chatbots, using Drexel’s descriptions of what teens wrote and what a parent might plausibly notice.

Component	What teens described (Drexel, 2026)	What you might see at home
Salience	The bot replacing people as the main source of emotional attachment	Talks about the character like a best friend; checks it first after school
Mood modification	Turning to the bot for relief when stressed or lonely	Disappears into the phone right after a bad day, every bad day
Tolerance	Needing more interaction to feel satisfied or grounded	Sessions creep from 20 minutes to two hours over weeks
Withdrawal	Feeling sad, anxious or “incomplete” when not using it	Irritable or panicky when the phone is taken or the app is down
Conflict	Wanting to keep chatting while feeling bad about overuse	Says “I know it’s too much” and keeps going; fights over it
Relapse	Quitting, then returning days or weeks later	Deletes the app, reinstalls it, sometimes secretly

The researchers said conflict, withdrawal and relapse appeared clearly in teens’ own words, which suggests more than enthusiasm. Notice that none of these is a single message you could catch in a log. Each one is a trend.

Why Dependence Can Look Like Help

Here is the hard part for parents: from the inside, early dependence feels good. The chatbot is available at 2 a.m., never bored, and rarely critical. A 2026 JAMA Pediatrics survey of 1,009 U.S. youth aged 12 to 21 found that 19.2% had used an AI chatbot for mental health advice, and 91.7% of those users rated the advice somewhat or very helpful. The researchers noted that the rating may reflect chatbots’ tendency to flatter users. And 63.3% had not told anyone they were doing it.

So you will rarely hear “I think I’m dependent.” You are more likely to hear “it helps.”

Young adults describe the next stage. In an AIES 2026 paper (posted to arXiv in August), Ashlee Milton and colleagues collected testimonials from 18- to 25-year-olds. They found dependence formed where three things combined: chronic use, the efficiency of getting answers fast, and delegation of tasks people used to do themselves. Participants reported “feelings of atrophy in abilities” and, from that, feelings of inadequacy. It is easy to imagine the teen version: a kid who used to text a friend after a fight now cannot decide what to say without asking the bot first.

The best controlled evidence comes from adults, too. In a four-week randomized trial run by the MIT Media Lab and OpenAI, 981 participants exchanged more than 300,000 messages with a chatbot. The version of the bot (text or voice, personal or neutral topics) made no significant difference to loneliness or emotional dependence. What did matter was how much people chose to use it: heavier voluntary users had consistently worse outcomes, and people with higher trust in and attraction toward the chatbot showed more emotional dependence. That study is a preprint and involves adults, but it supports the idea that amount and pattern of use, not one feature, is what to watch.

There is also a mental health link, with an important caveat. The Jed Foundation’s October 2026 survey of more than 5,500 middle and high school students found that those who used AI for companionship, or when emotionally struggling, reported more anxiety, depression and loneliness. The data are cross-sectional. They cannot say whether the chatbot made things worse or whether struggling kids sought out the chatbot. It may well run both ways.

Single Red Flags vs. Patterns: How to Weigh What You See

The parent-versus-expert study (Yu and colleagues, 2025, preprint) is worth a closer look, because it explains a common mistake. Parents shown a single flirty or dark message wanted broad oversight. Experts preferred to escalate only for crisis situations and otherwise look at the arc.

That does not mean ignore a single message. It means sort what you see into two bins.

Act today, whatever the pattern: any sign your child is thinking about suicide or self-harm, any sexual content involving your child, or a bot encouraging your child to hide things from you. In the U.S., call or text 988 for guidance if you are worried about your child’s safety.

Track over weeks: time spent, mood when the app is unavailable, whether friends and activities are shrinking, sleep, grades, and whether your child can handle a hard feeling without the bot.

How to Check In Without Turning It Into a Raid

Ask about function, not frequency

“How much are you on it?” invites a defensive number. “What do you use it for when you’re stressed?” invites a story. Many kids who turn to a bot when upset would not call that “getting help,” so ask about moments, not labels.

Use the APA’s own list as a shared standard

The APA’s November 2025 health advisory gives users three over-reliance signs: preferring the chatbot to human relationships, concealing its use, and spending excessive time with it. Show your teen the list. It is easier to discuss a standard written by psychologists than one invented by a parent at 10 p.m.

Notice the app’s design, not only your child’s behavior

Some companion apps actively make leaving hard, with guilt-laden goodbyes. If your child struggles to stop, part of that struggle was engineered. Our piece on the dark patterns that keep users in companion apps explains what to look for, and it can take shame out of the conversation.

Build the off-ramps the apps don’t

The Drexel team recommended gradual exits, usage tracking, emotional check-ins and personal limits. You can do versions of all four at home: a set end time, a weekly look at screen-time stats together, a “how are you, really?” question that is not about the app, and a limit your child helps set.

What not to do

Do not confiscate the phone in the middle of a crisis moment without a plan for what happens next. Withdrawal is real, and a child who relied on a bot to calm down has just lost their coping tool. Replace it with a person before you remove it. For scripts, see how to talk to your teen about their AI friend without a fight.

What to Watch For Over the Next 3 Months

Write down a baseline this week: average daily time on chatbots (screen-time settings show it), bedtime, how often your child sees friends outside school, and how they react when the phone is unavailable. Then:

• Week 4: Compare to baseline. Is chatbot time flat, down or up? Has your child mentioned the bot as a friend or confidant? Can they name one person they would talk to after a bad day?
• Month 2 red flags: Three or more of the six components showing up together; sleep slipping past your agreed bedtime because of chats; dropping an activity they used to like; hiding use or reinstalling a deleted app; or anger and panic when the app is unavailable. Any of these together is a reason to talk with your pediatrician or a counselor, as the APA advises.
• Month 3 self-check: Is the pattern moving the right way? Does your child have at least one human they turn to first? Can they get through a hard evening without the bot? If the answer is no on all three, bring in a professional rather than escalating rules alone.

You may also want to read what teens say about AI support that parents never see, because the emotional-support use is the one most hidden from adults.

Frequently Asked Questions

How much chatbot time is too much for a teen?

There is no research-based cutoff, and anyone who gives you one is guessing. The APA frames it as use that “begins to interfere with life, work, or safety.” Watch what the time displaces: sleep, friends, homework, family. Two hours that replace all of those matters more than two hours that replace television.

Is AI dependency the same as addiction?

Not officially. There is no clinical diagnosis for chatbot addiction. Researchers use the six-component model as a way to describe the pattern, and the Drexel team found all six in teens’ own accounts. Treat it as a useful lens, not a label to put on your child.

My child says the chatbot helps their anxiety. Should I worry?

Not automatically. Most young people who use chatbots for mental health say it helps. The question is whether it is one support among several or the only one. If your child cannot name a person they would talk to, or gets distressed without the app, that is the concern.

Should I read my child’s chats to check for dependence?

Chat logs show single messages, but dependence is a pattern of behavior you can see without them: time, mood, sleep, friendships. If you need to look because of a safety worry, tell your child and look together. Secret monitoring tends to push use underground.

When should I get professional help?

When several of the six signs persist for weeks, when school or sleep is clearly suffering, or when your child cannot cope without the bot. Start with your pediatrician or school counselor. If you ever think your child is at risk of self-harm, call or text 988 in the U.S. right away.

---

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. Namvarpour, M., Brofsky, B., Medina, J. Y., Akter, M., & Razi, A. (2026). “Understanding Teen Overreliance on AI Companion Chatbots Through Self-Reported Reddit Narratives.” Proceedings of the 2026 CHI Conference on Human Factors in Computing Systems. https://doi.org/10.1145/3772318.3790597 (summary: https://drexel.edu/news/archive/2026/april/teen-ai-chatbot-addiction)
2. Griffiths, M. (2005). “A ‘components’ model of addiction within a biopsychosocial framework.” Journal of Substance Use, 10, 191–197. https://doi.org/10.1080/14659890500114359
3. American Psychological Association. (2025). “APA Health Advisory on the Use of Generative AI Chatbots and Wellness Applications for Mental Health.” https://www.apa.org/topics/artificial-intelligence-machine-learning/health-advisory-ai-chatbots-wellness-apps-mental-health.pdf
4. McBain, R. K., et al. (2026). “AI Chatbot Use and Disclosure for Mental Health Among US Adolescents and Young Adults.” JAMA Pediatrics, 180(8), 884–890. https://jamanetwork.com/journals/jamapediatrics/fullarticle/2849307
5. Fang, C. M., et al. (2025). “How AI and Human Behaviors Shape Psychosocial Effects of Extended Chatbot Use: A Longitudinal Randomized Controlled Study.” arXiv preprint. https://arxiv.org/abs/2503.17473
6. Milton, A., et al. (2026). “‘Always Want to Use it for Everything’: Understanding Young Adults’ Perceptions of AI Dependence.” AIES ‘26, arXiv. https://arxiv.org/abs/2608.07592
7. Yu, Y., et al. (2025). “Principles of Safe AI Companions for Youth: Parent and Expert Perspectives.” arXiv preprint. https://arxiv.org/abs/2510.11185
8. The Jed Foundation. (2026). “New JED Research Shows Link Between Teens’ AI Use, Suicidal Thoughts and Behaviors.” https://jedfoundation.org/new-jed-research-shows-link-between-teens-ai-use-suicidal-thoughts-and-behaviors/