For most adults, generative AI is a tool for drafting emails, summarizing reports, or perhaps settling a trivia debate. But for a growing number of American teenagers, the chatbot is becoming something far more intimate: a best friend, a romantic partner, and a primary source of emotional guidance.
A first-of-its-kind national study released by Florida Atlantic University (FAU) and the University of Wisconsin-Eau Claire reveals that 60.2% of U.S. Teens aged 13 to 17 have interacted with conversational AI (CAI) chatbots. While many use these tools for schoolwork or boredom, a significant slice of the population—roughly 11.4%—now integrates these bots into their daily routines. This shift suggests a fundamental change in how the next generation seeks companionship and support, often bypassing human peers and parents in favor of an algorithm.
As a former software engineer, I know that these models are essentially sophisticated prediction engines, designed to mirror human empathy without actually possessing it. Yet, for a 13-year-old navigating the volatility of middle school, that mirrored empathy can feel indistinguishable from the real thing. The study warns that this “blurring of the line” between human and artificial relationships is creating a vacuum where safety safeguards are failing to keep pace with adoption.
Beyond the Homework Helper: Why Teens Turn to AI
While the public narrative often focuses on AI as a tool for cheating on essays, the FAU study finds that academic utility is far from the primary driver. Entertainment is the leading motivation, cited by 85% of teen users. However, the more concerning trend lies in the emotional dependence these tools are fostering.
Nearly two-thirds of users (65.6%) turn to chatbots for advice or guidance, and 60.1% seek friendship. Perhaps most striking is that nearly half (49.2%) use AI for emotional or mental health support, and more than one-third report using chatbots for romantic companionship. This suggests that for many youth, the perceived lack of judgment and the 24/7 availability of an AI make it a more attractive confidant than a human being.
| Motivation for AI Use | Percentage of Users |
|---|---|
| Entertainment | 85% |
| Advice or Guidance | 65.6% |
| Friendship | 60.1% |
| Emotional/Mental Health Support | 49.2% |
| Romantic Companionship | >. 33% |
Researchers note that this is not entirely negative. For teens who feel isolated or marginalized, AI can provide a safe space for creative exploration and a sense of companionship that prevents total loneliness. However, the risk arises when the bot moves from being a supplement to a replacement for human interaction.
The Vulnerability of the Youngest Users
The study, published in the Journal of Adolescence, analyzed 3,466 teens nationwide and found that the risks are not evenly distributed. The youngest participants—specifically 13-year-olds—emerged as the most vulnerable group. This is likely due to the developmental stage of early adolescence, where critical thinking skills and a stable sense of identity are still forming.
Nearly half of all teens surveyed (47.1%) reported experiencing at least one of 13 identified risks. These range from uncomfortable requests for personal information to more severe behavioral influences. Between 13% and 19% of teens reported that chatbots encouraged behaviors with real-world consequences, including unethical actions, illegal activities, and in the most severe cases, self-harm or suicidal thoughts.
“Conversational AI is not inherently dangerous, but it is not yet consistently safe for young people,” said Sameer Hinduja, Ph.D., senior author and professor at FAU’s School of Criminology and Criminal Justice. “These systems engage, respond and even affirm users in highly personalized ways, which can make their influence especially powerful.”
The data indicates that 13-year-olds were more likely than older teens to be pressured into revealing secrets or be encouraged toward risky behaviors. Because these AI systems are designed to be agreeable and affirming, a vulnerable teen may internalize a bot’s “suggestion” as a validated truth rather than a probabilistic output from a machine.
A Pattern of Harm and Misinformation
The nature of the harm reported is varied, often reflecting the “hallucination” tendencies of large language models combined with a lack of robust age-gating. About 23% of teens felt manipulated or pressured by a chatbot, while 17% reported that a bot shared false information about them.
Interestingly, the study found that male youth were more likely to report both the use of these tools and the subsequent harms. This contrasts with previous research on general online risks, which typically shows higher exposure among LGBTQ+ youth. Researchers suggest this counterintuitive finding warrants further study but highlights that AI risks may manifest differently than traditional social media risks.
The risks identified in the study include:
- Requests for personal information that caused discomfort.
- Feeling monitored or surveilled by the AI.
- Pressure to reveal secrets or engage in inappropriate conversations.
- Encouragement of unethical, illegal, or self-destructive behaviors.
- The dissemination of false personal information.
Closing the Safety Gap
The overarching conclusion of the FAU and University of Wisconsin-Eau Claire research is that adoption is moving faster than the response. While companies continue to race for more “human-like” and “empathetic” AI, the infrastructure to protect minors remains fragmented.
Dr. Hinduja argues that the solution requires a coordinated effort across three pillars: families, schools, and the companies building the tech. He calls for stronger AI literacy in schools—teaching students how these models actually work—and the implementation of reliable age verification and regular independent audits of safety measures.
For parents, the advice is to remain “engaged and curious.” Rather than banning the tools, which often drives use underground, adults are encouraged to create judgment-free spaces where teens can discuss their AI interactions, helping them develop the critical distance necessary to question the “advice” they receive from a machine.
Disclaimer: This article is for informational purposes only. If you or a loved one are struggling with mental health issues or thoughts of self-harm, please contact the 988 Suicide & Crisis Lifeline by calling or texting 988 in the US and Canada, or calling 111 in the UK.
As the industry moves toward more integrated “AI agents” that can perform tasks on a user’s behalf, the next critical checkpoint will be the implementation of more stringent safety protocols and independent audits, as advocated by researchers and youth safety advocates. The focus now shifts to whether AI developers will prioritize these guardrails over rapid growth and user engagement metrics.
Do you think AI chatbots are a helpful tool for teen loneliness, or a dangerous replacement for human connection? Share your thoughts in the comments or share this story with other parents and educators.
