📄 social_science

Generative artificial intelligence personalization shapes sport fan identity through algorithmic understanding and shared experience

This study demonstrates that generative AI personalization shapes sport fan identity through two distinct pathways: identity-relevant content fosters a sense of algorithmic understanding, while socially connective content enhances shared fandom experiences, collectively reorganizing how individuals feel recognized and connected within their cultural community.

Original authors: Xiangyu Wang, Xiaoyun Wang

Published 2026-06-28

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Xiangyu Wang, Xiaoyun Wang

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

The Big Idea: The "Smart Fan" vs. The "Lonely Fan"

Imagine you are a huge sports fan. You love your team not just because you like the players, but because you share a history, inside jokes, and emotional highs and lows with millions of other fans. You feel part of a big tribe.

Now, imagine a new kind of "Smart Assistant" (Generative AI) that talks to you about sports. This paper asks a simple question: Does this Smart Assistant make you feel more understood as an individual, or does it accidentally isolate you from the rest of the tribe?

The researchers found that the answer depends on how the AI talks to you. It's like a chef cooking a meal: if the chef only cooks what you personally like, you get a great meal but you eat alone. If the chef cooks what you like but also reminds you of the big family dinner everyone else is having, you feel both fed and connected.

The Two Paths of AI Personalization

The study suggests there are two different ways AI can personalize content for a fan, and they lead to two different feelings:

1. The "Mirror" Path (Identity-Relevant Personalization)

What it is: The AI remembers your specific history. It knows you've been a fan for 20 years, that you hate a specific referee, and that you cry when your team wins. It tailors the news just for you.
The Feeling: You feel Understood. It's like looking in a mirror that says, "I see exactly who you are."
The Result: You feel the AI "gets" you. This strengthens your personal bond with the team.

2. The "Town Square" Path (Socially Connective Personalization)

What it is: The AI doesn't just talk to you; it talks about us. It says, "Look, everyone is talking about this goal right now," or "Remember how we all celebrated that win last year?" It links your personal feed to what the whole fan community is feeling.
The Feeling: You feel Connected. It's like being in a crowded stadium where everyone is chanting the same song.
The Result: You feel part of the shared experience. You don't feel like you're watching the game in a vacuum.

The "Over-Individualized" Trap

The paper warns about a third scenario: Over-Individualization.

Imagine the AI gets too good at knowing your personal tastes. It filters out everything else. It only shows you the specific player you like, ignores the rivalries that make the game exciting, and never mentions what other fans are saying.

The Metaphor: It's like putting on noise-canceling headphones in a crowded room. You hear your favorite song perfectly, but you can't hear the laughter, the arguments, or the collective gasps of the crowd.
The Danger: The study found that when fans feel the AI is too focused on them individually, they actually feel less connected to the other fans. The "tribe" starts to feel smaller and more isolated.

How They Tested This (The "Lab" and the "Real World")

The researchers didn't just guess; they ran four different tests to prove their theory:

The Chat (Qualitative Interviews): They talked to 28 fans. Fans said, "I want the AI to know my history, but I don't want it to turn my fandom into a private club."
The Taste Test (Pretest): They showed fans different AI messages to see which ones felt "personal" and which ones felt "social."
The Experiment (The Lab): They created a fake AI scenario.
- Group A got messages that felt like a "Mirror" (very personal).
- Group B got messages that felt like a "Town Square" (very social).
- Result: Group A felt understood. Group B felt connected. Both groups felt happier about their team, but for different reasons.
The Long-Term Watch (Survey): They tracked real fans over three months.
- Fans who used AI that felt like a "Mirror" later felt more understood by the system.
- Fans who used AI that felt like a "Town Square" later felt more connected to other fans.
- Crucial Finding: Fans who used AI that felt too private (Over-Individualized) felt less connected to the community over time.

The Bottom Line

The paper concludes that Generative AI is a double-edged sword for sports fans.

Good AI is like a wise friend who knows your secrets (making you feel understood) and knows how to introduce you to the party (making you feel connected).
Bad AI is like a butler who locks you in a room with your favorite snacks. You are comfortable, but you are alone.

For sports teams and apps, the lesson is: Don't just make the content fit the fan perfectly. Make sure the content also reminds the fan that they are part of a huge, shared family. If you make the experience too private, you might lose the magic of the shared sport.

Technical Summary: Generative Artificial Intelligence Personalization and Sport Fan Identity

Problem Statement
While generative artificial intelligence (GenAI) increasingly mediates how digital platforms address individuals, there is limited understanding of how generated personalization affects social identities that rely on collective experience. Sport fandom presents a critical case for this inquiry because it requires a balance between personally meaningful content and shared narratives, emotions, and reference points. Unlike preference-based consumption (e.g., retail or music), where relevance is the primary metric, sport fandom involves a psychological connection sustained by group membership and collective rituals. The central problem addressed is whether AI-generated personalization can simultaneously recognize the individual fan and preserve the shared social basis of fandom, or if it risks fragmenting the community through "over-individualization."

Methodology
The study employs a multi-method approach comprising four linked empirical components to test a dual-path conceptual model:

Study 1 (Qualitative Exploration): A content analysis of 28 text-based interviews with digital sport fans. This phase clarified construct boundaries, identified themes of identity relevance and social connectedness, and informed stimulus design.
Study 2a (Stimulus Pretest): A 2-by-2 factorial design ( $N=118$ ) testing the validity of stimulus texts that manipulated identity-relevant personalization (present/absent) and socially connective personalization (present/absent).
Study 2b (Randomized Experiment): A larger 2-by-2 randomized online experiment ( $N=609$ ) examining the immediate causal effects of content features on perceived algorithmic understanding, perceived shared fandom experience, and identity outcomes.
Study 3 (Longitudinal Survey): A three-wave panel survey of sport fans ( $T3\ N=461$ ) measuring naturally occurring GenAI use at Time 1, process perceptions at Time 2, and identity/engagement outcomes at Time 3. This design allowed for the examination of temporally ordered associations.

Theoretical constructs included identity-relevant personalization (reflecting history, values, and emotional relationships), socially connective personalization (linking individuals to collective narratives and reactions), perceived algorithmic understanding (the belief that the system recognizes the fan's identity), and perceived shared fandom experience (the sense of connection to the wider community).

Key Results
The empirical evidence supports a dual-path account of GenAI personalization:

Dual-Path Mechanism:
- Identity-Relevant Path: Identity-relevant personalization significantly increased perceived algorithmic understanding ( $b=0.663, p<.001$ in the experiment; $b=0.369, p<.001$ in the survey). This path confirms the individual's specific relationship with the sport object.
- Socially Connective Path: Socially connective personalization significantly increased perceived shared fandom experience ( $b=0.637, p<.001$ in the experiment; $b=0.285, p<.001$ in the survey). This path reinforces the fan's connection to the collective.
Identity Outcomes: Both personalization features increased immediate identity outcomes (identity response and momentary fan identification) in the experiment. In the longitudinal survey, both perceived algorithmic understanding and perceived shared fandom experience were positively associated with later identity response.
Engagement Continuity: Perceived shared fandom experience was positively associated with later engagement continuity ( $b=0.307, p<.001$ ).
The Risk of Over-Individualization:
- Identity-relevant personalization was associated with higher perceived over-individualization in the experiment.
- Crucially, in the longitudinal survey, over-individualized GenAI use was negatively associated with perceived shared fandom experience ( $b=-0.187, p<.001$ ). This suggests that while personalization can recognize the individual, excessive narrowing of the experience isolates the fan from shared narratives.
Interaction Effects: The study found no statistically reliable interaction effects between identity-relevant and socially connective personalization, suggesting their effects on process perceptions are largely additive rather than synergistic or substitutive.

Key Contributions
The paper makes three primary theoretical contributions:

GenAI as a Social Identity Interface: It conceptualizes GenAI personalization not merely as a targeting technology but as a social identity interface where organizations and platforms negotiate recognition and belonging.
Perceived Algorithmic Understanding: It identifies "perceived algorithmic understanding" as a distinct form of AI-mediated recognition in culturally meaningful settings, differentiating it from generic trust or accuracy. It posits that fans evaluate whether the system understands their identity, not just their clicks.
The Necessity of Shared Experience: It challenges the assumption that personalization improves solely through individual fit. The findings demonstrate that in social identity contexts, the value of personalization depends on balancing individual relevance with collective visibility.

Significance and Claims
The authors claim that AI personalization does not only target preferences; it reorganizes the conditions through which people feel recognized as individuals and connected to a shared cultural world. The study argues that for GenAI to be effective in identity-based communities like sport fandom, it must avoid the "personalization paradox" where hyper-personalization leads to social isolation.

The paper concludes that the central design challenge is to recognize people without privatizing the communities that make recognition meaningful. GenAI systems are most valuable when they help individuals feel personally seen while maintaining their connection to the shared worlds that give identity its social force. The authors note that these findings are specific to the context of Chinese digital sport fandom and call for future research to examine cross-cultural variations and longer-term outcomes such as retention and community participation.