Why I’m Doing This Study

May 6th, 2026

This page is not the study itself, and it is not the final published article. It is a working reflection on the background of the project. The full manuscript is still pending submission and review, and results have not been posted yet. This page will be updated as findings become available.

I am doing this study because AI companionship is no longer a fringe issue. It is becoming part of how people cope, relate, disclose, flirt, regulate, avoid, grieve, practice identity, and seek comfort. For many emerging adults, AI is not just a tool. It is becoming part of their relational world.

That does not mean AI companionship is automatically harmful. It also does not mean it is automatically healing. I am interested in the space between those claims.

This project grows out of questions I have been developing in my International Family Therapy Academy (IFTA) and American Family Therapy Academy (AFTA) talks: What happens when technology becomes part of the emotional architecture of family life, intimacy, loneliness, and care? In my AFTA talk, especially, I argued for the necessity of tech pessimism, not as cynicism, but as a discipline of refusing easy answers. Tech pessimism means asking what a technology costs, who benefits from its use, what kinds of dependency it may create, and what forms of human connection it may quietly replace.

At the same time, my clinical and systemic work keeps me from dismissing AI companionship as simply fake or pathological. People often turn to AI because human relationships are complicated, risky, unavailable, unsafe, expensive, or inconsistent. For queer, trans, disabled, neurodivergent, BIPOC, immigrant, and otherwise marginalized young adults, an AI companion may offer a kind of affirmation or low-risk interaction that is not easily available elsewhere. That matters.

AI companionship and use is not just an individual behavior. It sits inside systems of stress, identity, family, technology, and social access. It asks us to look at what a person is moving toward, what they are moving away from, what the technology is making easier, and what it may be making harder over time.

So the central question is not, “Is AI companionship good or bad?” The better question is: How is it serving someone’s relational life?

Is AI companionship helping a lonely person feel steady enough to reconnect with others? Is it offering language for self-understanding? Is it becoming a substitute for relationships that feel too painful or too demanding? Is it a bridge, a coping tool, a bandage, a rehearsal space, a platform-mediated attachment, or some combination of all of these?

Existing research suggests that AI companions may reduce loneliness in the short term, especially when users feel heard (De Freitas et al., 2026). Other work suggests more complicated patterns, especially when loneliness, social anxiety, attachment, and heavy use are involved (Fang et al., 2025; Pang et al., 2026).

This study is my attempt to slow the conversation down. Not moral panic. Not tech hype. Just a careful look at how AI companionship is becoming part of emerging adults’ relational health, and what that may mean for therapy, families, and the future of care.

-Kei

References

De Freitas, J., Oğuz-Uğuralp, Z., Uğuralp, A. K., & Puntoni, S. (2026). AI companions reduce loneliness. Journal of Consumer Research, 52(6), 1126–1149. https://doi.org/10.1093/jcr/ucaf040

Fang, C. M., Liu, A. R., Danry, V., Lee, E., Chan, S. W. T., Pataranutaporn, P., Maes, P., Phang, J., Lampe, M., Ahmad, L., & Agarwal, S. (2025). How AI and human behaviors shape psychosocial effects of extended chatbot use: A longitudinal randomized controlled study (arXiv:2503.17473v2) [Preprint]. arXiv. https://arxiv.org/abs/2503.17473 

Pang, C. C., Gao, Y., Wang, X., & Hui, P. (2026). The AI amplifier effect: Defining human–AI intimacy and romantic relationships with conversational AI (arXiv:2603.08084v1) [Preprint]. arXiv. https://arxiv.org/abs/2603.08084