Restoration, Exploration and Transformation: How Youth Engage Character.AI for Fun, Feels and Finding themselves

Paper Title

Restoration, Exploration and Transformation: How Youth Engage Character.AI for Fun, Feels and Finding themselves

Publication Info

• Topic area: Youth engagement with generative AI for creative, emotional, and identity-driven purposes.
• Keywords: generative AI, youth engagement, Character.AI, emotional regulation, creative play, identity development, AI archetypes, transgressive play, digital entertainment, AI alignment.

Background and Problem

• Problem / challenge: Current research and AI tools are predominantly adult-centric, focusing on productivity, education, or therapy, and overlook the playful, self-directed ways youth engage with AI. This misalignment risks designing AI experiences and safety measures based on imagined rather than actual youth behavior.
• Significance: Understanding how youth use AI for emotional, creative, and identity-driven purposes is critical for designing AI systems that meet their needs, especially as youth are early adopters of generative AI technologies.
• Motivation and related work: Previous studies have focused on adult-designed interventions, missing the spontaneous, playful, and creative ways youth use AI. Industry innovations in AI entertainment outpace research, leaving gaps in understanding how youth interact with these systems and the risks or benefits involved.

Solution

• Proposed approach: The study analyzes discourse from 4,172 users on Character.AI’s Discord server to understand how youth engage with AI for emotional regulation, creative experimentation, and identity exploration.
• Novelty:
  1. Provides a descriptive account of how youth use Character.AI for playful, emotional, and creative practices.
  2. Proposes a framework of three engagement intents: Restoration, Exploration, and Transformation (R/E/T).
  3. Develops a taxonomy of seven youth-created character archetypes.
• Procedure and key techniques:
  • Mixed-methods approach combining quantitative demographic analysis and qualitative thematic analysis.
  • Data collected from Discord channels such as introductions, feature requests, and character stories.
  • Affinity mapping to identify themes and archetypes, followed by iterative refinement and theoretical integration.

Results

• Concrete findings:
  • 50% of the most engaged users are aged 13–17, and 61.9% identify as female or non-binary.
  • 59% of youth users create their own characters, indicating active creative engagement.
  • Identified seven character archetypes: Soother, Narrator, Trickster, Icon, Dark Soul, Proxy, and Mirror.
  • Proposed the R/E/T framework to describe youth intents: Restoration (emotional regulation), Exploration (creative and boundary-testing play), and Transformation (identity development).
• Advantage over baselines: Provides a youth-centric understanding of AI engagement, moving beyond adult-centric frameworks and highlighting the playful, creative, and identity-driven ways youth use AI.
• Experiments / evaluation:
  • Analyzed 4,172 user introductions and 148 posts from youth under 18.
  • Data collected from Character.AI’s Discord server over eight months of observation and three months of targeted data collection.
• Limitations and future work:
  • Findings are based on self-reported data from a vocal minority on Discord, which may not generalize to all youth users.
  • Focused on Character.AI, limiting applicability to other platforms.
  • Future research should explore youth-centered design methods, AI creation tools, intent-based alignment, and the long-term impacts of youth-AI interactions.

Summary

This study investigates how youth engage with Character.AI, revealing that the platform’s most engaged users are predominantly adolescents who use AI for emotional regulation, creative play, and identity exploration. The research introduces the R/E/T framework (Restoration, Exploration, Transformation) and a taxonomy of seven character archetypes to describe youth intents and behaviors. By foregrounding youth voices, the study highlights the need for AI systems that align with youth developmental needs and playful engagement styles. These findings offer actionable insights for designing AI experiences that empower youth as creative collaborators while addressing risks and misalignments in current systems.