As artificial intelligence chatbots become a normalized source of emotional support for adolescents, researchers from Arizona State University are raising concerns that widespread adoption could compromise the development of critical interpersonal skills during a formative developmental window. A recent analysis published in The Lancet Child & Adolescent Health warns that while AI-powered conversational tools offer immediate, nonjudgmental guidance, they may displace the emotionally complex interactions through which teenagers typically learn relationship competencies.
The concern arrives as adoption metrics show significant teen engagement with AI systems. Research cited by the ASU team found that 64% of U.S. adolescents now use interactive AI tools, with 42% having used AI chatbots for friendship-related advice and 19% for guidance on romantic relationships. Platforms such as ChatGPT, Replika, Claude, and Character.AI have become de facto counselors for teenagers navigating family conflict, peer dynamics, and relationship challenges-experiences that researchers argue should typically unfold through direct human interaction.
Thao Ha, lead author and associate professor in ASU’s Department of Psychology, framed the core issue as a developmental timing problem. Adolescence represents a critical period for acquiring emotional regulation, conflict resolution, perspective-taking, and boundary-setting skills. These competencies emerge through emotionally charged interactions with peers, romantic partners, and family members. “People don’t realize that relational learning happens during the teenage years and that these moments of social connection are little building blocks that become bigger things that will benefit you throughout life,” Ha said in the published work.
The Skills Gap Concern
The researchers note that AI systems, regardless of sophistication, lack the reciprocal emotional stakes of human relationships. When a teenager asks a chatbot for advice about a conflict with a friend, the interaction follows a predictable pattern: the bot offers guidance, the teen absorbs it, and the exchange ends. A real friendship navigates the same conflict through negotiation, apology, repair, and mutual vulnerability. The latter process is messy, risky, and irreplaceable in terms of skill development.
The ASU team consulted with youth advisory board members, including two high school students ages 16 and 17. One senior reflected on the breadth of teen AI adoption: “I don’t think I really expected for so many teens to have the same concerns or thoughts when it came to AI. We all mostly had concerns about how AI was replacing actual human connection and how it limits a lot of those needs that humans have that cannot be replaced with a computer artificial intelligence.” This observation from teenagers themselves suggests that awareness of the tradeoff exists among users, yet engagement continues.
Regulatory Design Lags Behind Adoption
Current safeguards prove inadequate. Teenagers told the ASU researchers that existing regulatory approaches, such as age verification, are ineffective and do not reflect their actual needs or usage patterns. The gap between technology rollout and governance is substantial. “The technologies are developing super-fast, faster than we can keep up with as scientists, faster than governance and policy can keep up with,” Ha noted in the published research.
This mismatch reflects a broader pattern in Digital Health and consumer technology. Digital health startups face complex regulatory environments encompassing FDA oversight, HIPAA, and emerging AI laws, yet even those frameworks remain under active construction. Consumer AI chatbots, which often lack clinical claims and operate outside traditional health regulation, face even less oversight.
The Upside and Caution
The ASU researchers acknowledge that AI chatbots offer genuine benefits. For teenagers in isolation, without trusted adults, or experiencing crisis, an available nonjudgmental digital resource can reduce harm. The concern is not that AI support exists, but that it may crowd out the developmental opportunities embedded in human relationships. The issue is one of balance, substitution, and design intentionality.
The authors call for safeguards and careful system design that positions AI as a supplement to human connection rather than a replacement. This might include features that encourage offline problem-solving, cap session frequency, or explicitly redirect certain categories of questions toward trusted adults. It could also mean transparency about the system’s limitations in a format that resonates with adolescent users.
The research does not propose banning teen access to AI tools. Instead, it argues for intentional policy design that preserves the developmental value of human relationships while allowing teenagers to benefit from immediate, affordable support when appropriate. Whether such design will occur faster than teen adoption is an open question. The data suggests that neither age verification nor parental control mechanisms have slowed uptake significantly. The challenge facing policymakers, platform designers, and parents is how to manage a technology that teenagers themselves recognize as both helpful and developmentally risky.
