"What Keeps Fans on the Silent Field?": Understanding Lean-Back Football Fans via AI Sports Broadcasting in Non-Event Time

Paper Title

'What Keeps Fans on the Silent Field?': Understanding Lean-Back Football Fans via AI Sports Broadcasting in Non-Event Time

Publication Info

• Topic area: Enhancing engagement for passive football viewers during uneventful match periods using AI-generated commentary.
• Keywords: Lean-back viewers, AI commentary, Non-Event Time, sports broadcasting, user-directed personalization, cognitive load, social presence, football fandom, HCI, generative AI.

Background and Problem

• Problem / challenge: Lean-back football fans struggle to stay engaged during Non-Event Time (NET) in matches, as current commentary options are either too sterile or cognitively demanding.
• Significance: Addressing this issue could improve the viewing experience for a significant audience segment, preserving their connection to live sports and enhancing media consumption.
• Motivation and related work: Existing solutions like official broadcasts and co-streaming fail to meet the needs of passive viewers. Prior research has focused on active engagement or peak moments, neglecting the experience during uneventful periods. This paper aims to explore how personalized AI commentary can sustain engagement during NET.

Solution

• Proposed approach: ARUA (Auditory Reaction & User-directed Articulation system), a prototype allowing users to preemptively direct AI-generated commentary tailored to their preferences.
• Novelty:
  1. Characterizes lean-back viewers as a distinct demographic seeking immersion without cognitive strain.
  2. Introduces a user-directed paradigm for creating personalized, low-effort media experiences.
  3. Provides design principles for AI systems that balance social presence, emotional tone, and cognitive load.
• Procedure and key techniques:
  • Users configure commentary teams using a creative interface, selecting personas (e.g., Caster, Analyst, Fans) and providing textual instructions.
  • The system synthesizes commentary using LLMs and TTS engines, integrating user inputs with contextual match data.
  • A qualitative study with 32 participants explored how users directed commentary for nine NET scenarios, analyzing their choices and motivations.

Results

• Concrete findings:
  • Heavy Fan personas were the most selected (40.1%), often combined with professional roles like Caster or Analyst.
  • Participants explored an average of 5.72 unique group-level combinations across trials, with high transition rates (77.0%).
  • Commentary was used as a relational tool, serving as emotional proxies, balanced panels, or intellectual counterpoints.
• Advantage over baselines:
  • ARUA addressed the double failure of official broadcasts (underload) and co-streaming (overload) by offering tailored, low-effort engagement.
  • Participants reported greater emotional resonance and entertainment compared to existing options.
• Experiments / evaluation:
  • Conducted with 32 football fans (30 males, 2 females) supporting English Premier League teams.
  • Participants directed commentary for nine randomized NET scenarios (3 scoreline contexts × 3 NET types).
  • Data collected included persona selections, textual instructions, and post-task interviews.
• Limitations and future work:
  • Fixed voices and isolated NET clips limited ecological validity; future studies should explore dynamic match contexts and vocal variations.
  • Gender imbalance in the sample (94% male) restricts generalizability; future work should recruit more diverse participants.
  • Potential extensions include embodied AI commentators in spatial computing environments.

Summary

This paper introduces ARUA, a prototype for personalized AI commentary aimed at enhancing engagement for lean-back football fans during Non-Event Time (NET). Through a qualitative study, the authors demonstrate how users craft commentary to balance emotional resonance, cognitive load, and social presence. Findings reveal that lean-back viewers use AI personas as relational tools, creating emotional proxies, balanced panels, and intellectual counterpoints. The study highlights the potential of user-directed AI systems to transform passive media consumption, offering actionable design principles for future applications in sports broadcasting and beyond.