Rehearsing the Self with AI: Teens Cross-Domain Use of Chatbots

Paper Title

Rehearsing the Self with AI: Teens Cross-Domain Use of Chatbots

Publication Info

• Topic area: Teen interactions with generative AI chatbots across academic, emotional, and social domains.
• Keywords: generative AI, chatbots, teens, identity formation, ambient trust, self-presentation, emotional regulation, algorithmic mediation, boundary regulation, HCI.

Background and Problem

• Problem / challenge: Existing research isolates single domains of chatbot use (e.g., academic, emotional support) and lacks empirical understanding of how teens integrate generative AI into their daily lives across multiple domains.
• Significance: Adolescence is a critical stage for identity exploration and emotional regulation, and the cross-domain use of AI chatbots may influence these developmental processes in subtle but significant ways.
• Motivation and related work: Prior studies have examined algorithmic mediation, emotional support, and identity performance in digital environments but have not addressed the cumulative effects of cross-domain chatbot use. This paper builds on theories of boundary regulation, context collapse, and algorithmic self-presentation to explore how teens navigate these interactions.

Solution

• Proposed approach: The concept of ambient trust is introduced to describe how repeated, emotionally neutral interactions with AI chatbots foster alignment and influence self-expression without deep emotional reliance.
• Novelty:
  1. A thematic account of teens’ hybrid, everyday chatbot practices across academic, emotional, and social domains.
  2. A theoretical synthesis drawing on boundary regulation, image management, and algorithmic mediation frameworks.
  3. The concept of ambient trust as a mechanism for understanding cumulative influence in emotionally neutral interactions.
  4. Design implications for developmentally appropriate and transparent AI systems.
• Procedure and key techniques: Semi-structured interviews with 20 U.S. teens aged 13–18 were analyzed using reflexive thematic analysis to identify patterns in how teens navigate and interpret cross-domain chatbot use.

Results

• Concrete findings:
  • Teens define “personal” narrowly as secret or risky, enabling emotionally meaningful interactions to feel safe and impersonal.
  • Instrumental convenience fosters habitual use, leading to a sense of alignment and familiarity (ambient trust).
  • Chatbots are used for tone adjustment, impression management, and identity rehearsal, often without teens recognizing these activities as identity work.
  • Teens express ambivalence toward chatbots’ relational features, appreciating responsiveness while rejecting overly human-like behaviors.
  • Teens maintain editorial control and enact boundary-setting strategies to preserve authorship and voice.
• Advantage over baselines: The study identifies ambient trust as a novel mechanism of influence, distinct from emotional reliance or parasocial intimacy, highlighting the cumulative effects of mundane interactions on identity shaping.
• Experiments / evaluation: Interviews explored teens’ experiences with chatbots across academic, emotional, and social domains, focusing on trust, privacy, and emotional impact. Participants were diverse in age, gender, and race/ethnicity.
• Limitations and future work: The sample is not nationally representative, and findings rely on self-reported data rather than trace analysis. Future work could examine longitudinal effects of cross-domain chatbot use and develop interventions to enhance algorithmic literacy among teens.

Summary

This paper investigates how U.S. teens aged 13–18 use generative AI chatbots across academic, emotional, and social domains, revealing nuanced practices of tone management, impression shaping, and identity rehearsal. It introduces the concept of ambient trust to explain how repeated, emotionally neutral interactions foster alignment and influence self-expression without deep emotional reliance. Findings highlight the developmental stakes of cross-domain chatbot use, including subtle identity shaping through cumulative personalization. Design implications emphasize transparency, boundary-setting, and algorithmic literacy to support teens’ autonomy and authorship in mediated environments.