Restoration, Exploration and Transformation: How Youth Engage Character.AI for Fun, Feels and Finding themselves
This is a strong CHI honorable-mention style contribution because it combines a clear empirical account with a memorable framework and a useful taxonomy. Its main value is not technical novelty in the system sense, but a well-supported reframing of youth AI use around lived practices, with clear implications for design and future research.
Axes Lens
Rare contribution shape, typical evidence profile. The point here is not a score. It is to show what kind of claim the paper makes, and whether the evidence pattern is unusual or baseline in this 268 -review set.
Contribution shape
- Knowledge form
- descriptive knowledge typical · 92/268
- Novelty type
- framework typical · 59/268
- Abstraction level
- practice typical · 85/268
- Generalization target
- user population typical · 75/268
- Validation mode
- qualitative study typical · 63/268
Evidence profile
- Evidence strength
- strong typical · 158/268
- Claim alignment
- strong typical · 231/268
- Overclaim risk
- medium typical · 210/268
Review Summary
This paper’s strongest contribution is conceptual rather than technical: it reframes youth engagement with Character.AI away from the usual adult-centered narratives of productivity, safety, or benchmark performance and toward a richer account of playful, emotional, and identity-oriented use. The abstract makes the contribution legible and the evidence basis is coherent: discourse from 4,172 users in the official Discord, with a focus on highly engaged youth, supports a descriptive account plus a three-part framework and a seven-archetype taxonomy. That combination is exactly the kind of CHI contribution that can travel well, because it gives researchers and practitioners a vocabulary for discussing what youth are actually doing with generative AI. At the same time, the paper’s own scope matters a lot. The cohort is self-selected, highly engaged, and drawn from a platform community where disclosure and performance are part of the interaction. That makes the findings valuable as a study of a specific youth-led practice community, but not as a general statement about all youth or all Character.AI use. So the paper is best read as a strong descriptive and framework-building contribution with medium overclaim risk if its categories are treated as universal. In field terms, it is important because it surfaces a mismatch between current AI experiences and youth practices, and it does so with enough empirical grounding to justify the design implications, while still leaving room for follow-up work on broader populations and in-product behavior.
What Changed
Canon before
Prior CHI work on youth and generative AI has largely emphasized adult-designed tools, controlled experiments, or productivity-oriented uses rather than self-directed, playful, emotionally expressive youth practices in platform-native communities.
Departure from common sense
The paper’s core departure is that youth are not framed as passive recipients of AI or as users mainly seeking utility; instead, the authors argue that adolescents actively repurpose Character.AI for playful, emotional, and identity-related ends, pushing platform limits and challenging adult-centered assumptions about what youth AI use looks like.
Actual novelty
The paper’s novelty is a combined contribution: a descriptive account of highly engaged youth discourse on Character.AI’s Discord, a three-part engagement framework (Restoration, Exploration, Transformation), and a taxonomy of seven youth-created character archetypes. The paper presents these as new analytic lenses for understanding youth AI use and for informing youth-centred AI design.
Evidence
The paper grounds its claims in naturalistic analysis of discourse from 4,172 users in Character.AI’s official Discord, with emphasis on highly engaged youth. The abstract explicitly states the three-part contribution and the platform-specific, youth-centered framing. The evidence supports a descriptive and framework-building contribution, but the scope remains bounded to a self-selected Discord cohort.
“ Information & Contributors Bibliometrics & Citations Reading Options References Figures Tables Media Share Abstract Young people are among the fastest adopters of generative AI, yet research emphasises adult-desig”
actual novelty · 007_share-on / Abstract + contributions statement · confidence 0.92
“ Cited By View all AI Generated (2026) Session Summary Podcast: Session 143: Mental Wellbeing Proceedings of the 2026 CHI Conference on Human Facto”
departure from common sense · 007_share-on / 1 Introduction / Abstract framing · confidence 0.78
“AI is a popular product broadly focused on AI entertainment, and the official Discord community is one of the few accessible spaces where adolescents organically discuss their self-directed AI use with peers[ 67 ].”
limitation · 007_share-on / 7.3 Limitations and Open Questions · confidence 0.90
“ To understand the characteristics of CAI’s most engaged users on Discord, we analysed 4,172 self-introductions from its official Discord serve”
validation scope · 007_share-on / 3 Thematic analysis approach + 4 Results + Discussion scope · confidence 0.74
Limits
Method limits
The study relies on discourse from a self-selected, highly engaged Discord cohort, so the analysis is shaped by who chose to speak and disclose. Identities are not independently verified, and the paper’s own framing suggests the findings are descriptive rather than causal or representative of all youth AI users.
Deployment limits
The framework and archetypes are most directly applicable to platform-native youth communities and to design/research contexts concerned with emotionally expressive, playful, or identity-oriented AI use. They should not be treated as a universal model of all youth interaction with generative AI or as a direct proxy for in-product behavior across platforms.
Boundary conditions
Findings are bounded by the Character.AI Discord setting, by highly engaged users, and by voluntary self-disclosure. Generalization is strongest for similar youth-led, community-discussion contexts and weaker for silent users, less engaged users, other platforms, or settings where identity and intent cannot be inferred from discourse.
Position in field
This paper sits at the intersection of youth studies, generative AI, and HCI design research. It advances the field by shifting attention from adult-centered evaluations of AI toward youth self-directed meaning-making, and by offering a vocabulary for emotional regulation, creative experimentation, and identity development in AI use.