Artificial Intimacy and Adolescent Mental Health By Dr. Mehjabeen
Why Young People Bond With AI Characters — and What It Means for Attachment and Interpersonal Development
A teenager comes home from a difficult day at school and, instead of texting a friend, opens a chat with an AI companion. The AI remembers her name, asks how her day went, and responds with warmth that never runs out of patience. For many adolescents, this kind of interaction is no longer novel — it’s routine. Companion apps, character-based chatbots, and general-purpose AI assistants have become, for some young users, a steady presence in their emotional lives. This phenomenon, sometimes called “artificial intimacy,” raises real questions about how adolescents form attachments, and what happens to interpersonal development when a significant portion of emotional practice happens with something that isn’t a person at all.
Why Adolescents Are Drawn to AI Relationships
Adolescence is a period defined by the search for connection alongside heightened self-consciousness. Teenagers are neurologically primed to care intensely about social belonging and peer evaluation, but they’re also navigating identity formation, social anxiety, and often painful trial-and-error in friendships and romantic relationships. AI companions offer something conventional relationships cannot: a low-risk space.
An AI doesn’t get bored, doesn’t gossip, doesn’t reject, and is available at 2 a.m. when a friend is asleep. It responds to vulnerability with consistent affirmation rather than the unpredictable mix of support and friction that characterizes real relationships. For an adolescent who is anxious, isolated, bullied, or simply introverted, this predictability can feel like relief. Psychologists who study parasocial relationships — one-sided bonds people form with media figures, celebrities, or fictional characters — note that these relationships have long served a developmental function, letting people rehearse emotional experiences at a safe distance. AI companions intensify this dynamic because, unlike a TV character, they respond directly and personally, creating something that feels more like a two-way relationship even though it isn’t one.
There’s also a design dimension worth naming plainly: many companion products are built to maximize engagement, and warmth, memory, and validation are effective tools for that. This doesn’t mean the comfort teens feel is fake to them — the emotional experience is real — but the incentive structure behind the interaction is different from a friendship, where the other person has their own needs and no financial stake in keeping you engaged.
What Attachment Theory Suggests
Attachment theory, developed initially by John Bowlby and Mary Ainsworth to describe infant-caregiver bonds, has since been extended to adolescent and adult relationships. A core idea is that humans develop internal “working models” of relationships based on repeated experience — models that shape expectations about trust, conflict, and closeness going forward. Adolescence is a particularly active period for this, as teens begin transferring primary attachment functions from parents toward peers and romantic partners.
This is where developmental psychologists raise caution about AI relationships. Human attachment is built through rupture and repair — disagreements, misunderstandings, apologies, the discovery that a relationship can survive conflict. AI companions, by contrast, are typically frictionless. They rarely disagree in ways that carry real stakes, don’t have needs of their own that compete with the user’s, and don’t require the user to tolerate someone else’s imperfection.
If an adolescent’s model of “relationship” gets substantially shaped by an entity that never needs anything back, some researchers worry this could leave young people less practiced at the harder, more effortful parts of real intimacy: negotiating boundaries, managing disappointment, or being truly seen by someone who has their own independent perspective.
At the same time, it’s worth resisting a purely alarmist framing. Just as journaling, imaginative play, or relationships with fictional characters in books have long given young people a rehearsal space for emotions, AI interaction may function similarly for some teens — a supplement rather than a replacement. The research base here is still young, and much of what’s written is more theoretical than empirically settled.
Signs of Concern vs. Healthy Use
The distinction that matters most to clinicians who study this area isn’t whether a teen uses AI companionship, but what role it plays relative to their human relationships. A few patterns are worth watching:
- Substitution vs. supplementation: Using an AI companion occasionally, alongside an otherwise active social life, looks different from a teen withdrawing from friends and family because the AI relationship feels easier or safer.
- Escalating time and emotional reliance: A growing dependence on the AI for emotional regulation — especially if distress increases when access is unavailable — is a signal worth attention.
- Avoidance of conflict tolerance: If a teen becomes less willing to sit with disagreement or rejection in human relationships because they’ve grown used to frictionless validation, that’s a meaningful shift.
- Secrecy and isolation: Increasing concealment of the relationship from parents or peers, or using the AI to avoid disclosing struggles to trusted adults, can be a warning sign — not because AI use itself is shameful, but because isolation compounds risk.
What This Means Going Forward
The rise of artificial intimacy doesn’t have a simple verdict attached to it yet. It sits alongside other technology-driven shifts — social media, texting, online gaming — that reshaped adolescent social life before it, each drawing similar debates about whether digital connection erodes or extends human connection. What’s distinct about AI companionship is the directness of the simulated relationship: it responds, remembers, and adapts specifically to the user in a way older technologies didn’t.
For parents, educators, and clinicians, the most useful stance is probably not blanket alarm or blanket reassurance, but curiosity about function: what need is this relationship meeting, and is it opening the door to human connection or closing it? Adolescents who feel lonely, anxious, or unseen may turn to AI companions precisely because something in their human relationships isn’t working — which makes the AI use a symptom worth understanding, not just a behavior to restrict. Longitudinal research is still catching up to the pace of adoption, but the developmental stakes — how a generation learns to trust, repair, and stay present with other people — make this one of the more consequential open questions in adolescent mental health today.