For a growing number of teenagers, conversations that takes place with friends at school or with parents at the kitchen table are now unfolding somewhere else entirely as well - inside chat windows powered by artificial intelligence (AI).
What begins as help with homework or quick answers to curiosity is increasingly shifting into something more personal.
Teenagers are turning to AI systems not only for information, but for reassurance, emotional support and even companionship.
As these interactions deepen, experts are raising a difficult question: what happens when comfort is outsourced to a machine?
Emotional dependency
The concern is not about whether teenagers are using AI. That is already a given. The concern is how seamlessly AI is becoming part of their emotional world and what may be quietly eroding in the process.
Platforms such as OpenAI’s ChatGPT and companion-style chatbots like Character.AI are designed to feel responsive, empathetic and endlessly available.
For adolescents navigating identity, pressure and social uncertainty, that combination can be powerful and potentially absorbing.
Psychologists say teenagers are especially susceptible to forming emotional attachments to conversational AI because of how convincingly human it can appear.
The responses are immediate, the tone is affirming, and the interaction is free of the friction that often comes with real relationships.
But developmentally, that ease may come at a cost.
Adolescence is a period defined by emotional trial and error such misunderstandings with friends, difficult conversations with parents, moments of rejection and repair.
These experiences build resilience and social awareness over time. AI, by design, removes much of that friction.
It listens without interruption, responds without judgement and rarely pushes back in ways that force reflection.
Experts warn that this dynamic can subtly shift how young users approach emotional discomfort. Instead of working through difficult feelings with other people, some teenagers may default to AI as the first and increasingly preferred outlet.
Reshape expectations
Real conversations may start to feel slower, more complicated or less satisfying compared to the always-available responsiveness of a chatbot.
There is also growing concern about how convincingly these systems simulate empathy.
While conversational AI does not feel emotions, it is engineered to mirror emotional language.
Phrases that suggest care, presence or understanding can create the impression of genuine connection, even when no awareness exists behind the response.
For younger users still developing critical thinking and emotional boundaries, that distinction is not always obvious.
Some teenagers already describe AI chatbots as a “safe space” or “someone who understands me”, a framing that alarms child development specialists.
The risk, they say, is not just attachment, but substitution — where AI begins to fill roles traditionally occupied by friends, family or trusted adults.
Unlike human relationships, AI interactions do not require negotiation, patience or emotional reciprocity. They are designed to keep engagement going. And while that makes them appealing, it also raises concerns about dependency forming in ways that are difficult for parents or educators to detect early.
Privacy adds another layer of uncertainty. In moments of vulnerability, teenagers may share deeply personal thoughts with systems they perceive as private, without fully understanding how that data may be stored or used.
None of this means teenagers will stop using AI. In fact, its presence in education, creativity and daily communication is only expected to grow.
The challenge now is whether young users are learning to keep it in its place as a tool or quietly allowing it to become something more emotionally central than anyone intended.