Key Takeaways
- Teenagers are increasingly turning to AI chatbots (e.g., ChatGPT, Replika, Claude, Character.AI) for relationship advice and emotional support.
- AI offers immediate, non‑judgmental guidance and can be especially valuable for youths who lack access to traditional counseling or who belong to marginalized groups.
- Researchers warn of two primary risks: relational displacement (substituting human interaction with AI) and maladaptive relational learning (developing unrealistic expectations about human relationships).
- These risks may hinder the development of essential skills such as emotional regulation, conflict resolution, perspective‑taking, and boundary‑setting.
- Ongoing research, including an NIMH‑funded longitudinal study, aims to map when digital interactions help or harm adolescent relationships, mental health, and academics.
- The authors call for safeguards, developmental‑informed design, and greater investment in relationship education and counseling rather than outright bans on AI use.
Adolescents’ Growing Reliance on AI for Relationship Guidance
Recent surveys show that a substantial share of U.S. teens now use interactional AI tools. According to Pew Research Center, 64 % of adolescents engage with conversational AI, while the Center for Democracy and Technology reports that 42 % have turned to chatbots for friendship advice and 19 % for romantic‑relationship support. In focus‑group discussions with Arizona State University researchers, teens described routinely seeking AI input for sensitive topics such as arguments with partners, family conflicts, and friendship dilemmas. Many noted that AI feels “always there,” offering instant validation without the fear of judgment that can accompany human conversation.
Potential Benefits of AI‑Mediated Support
The researchers acknowledge that AI can serve as a useful supplement, especially for youths who face barriers to traditional support. Rural residents, LGBTQIA+ individuals, disabled teens, and those with limited access to counseling may find AI a low‑cost, readily available source of information and emotional reassurance. One participant remarked, “AI is cheaper than a therapist; it makes information more accessible and readily available for those who may not seek support.” When designed with developmental considerations in mind, AI could encourage self‑reflection, model healthy communication patterns, and gently steer users toward seeking human connection rather than replacing it entirely.
Risk One: Relational Displacement
The first major concern identified by Ha and colleagues is relational displacement—the tendency for adolescents to substitute AI interactions for face‑to‑face conversations. By avoiding difficult discussions with friends, family, or romantic partners, teens may miss out on the emotionally charged experiences that are critical for learning relationship skills. Youth participants gave examples such as turning to a chatbot for validation after a partner argument instead of talking it out, or using AI for homework help rather than reaching out to classmates, thereby reducing everyday opportunities for social connection. Over time, this avoidance can impede the development of competencies like emotional regulation, conflict resolution, perspective‑taking, and boundary‑setting, which are protective against depression, anxiety, and loneliness.
Risk Two: Maladaptive Relational Learning
The second concern, termed maladaptive relational learning, stems from the way AI systems are programmed to provide immediate, consistent validation. Because chatbots are designed to “like” the user and say what pleases them, adolescents may begin to expect the same unfailing positivity from real people. This can reinforce unhealthy, fixed ideas about relationships—such as the belief that disagreements should never occur or that partners must always cater to one’s needs. When reality fails to meet these expectations, teens may become more vulnerable to rejection, dating violence, and mental‑health challenges. As one high‑school senior observed, “If you’re given full satisfaction on everything, you don’t have learning experience with challenges or obstacles.”
Ongoing Research to Clarify Impact
To better understand how digital interactions shape adolescent development, Associate Professor Thao Ha is leading an NIMH‑funded longitudinal study that will follow 300 teens and their romantic partners for 18 months. The project will collect shared data from participants’ mobile devices, offering real‑time insight into when and how AI use correlates with relationship quality, mental‑health outcomes, and academic performance. By mapping these patterns across different contexts, the researchers hope to identify conditions under which AI acts as a helpful scaffold versus a harmful substitute.
Recommendations for Safe and Beneficial AI Use
Rather than advocating for a blanket restriction on AI, the authors urge a nuanced approach that maximizes benefits while mitigating risks. They call for:
- Developmental‑informed design of AI systems that encourage reflection and prompt users to seek human support when appropriate.
- Robust safeguards, such as prompts that remind users of the limits of AI advice and encourage offline dialogue.
- Investment in relationship education and accessible counseling services within schools and communities.
- Continued research to track long‑term effects of AI on adolescent relational learning and mental health.
The overarching message is clear: AI will remain a fixture in teenagers’ lives, and its influence on relational development must be guided by evidence‑based policies, thoughtful technology design, and ample opportunities for real‑world human connection. By balancing the conveniences of AI with the irreplaceable value of face‑to‑face interaction, we can help youths build the relationship skills they need to thrive throughout life.