AI Sports Broadcasting Wins Over Viewers When Ease of Use Comes First, Study Finds

Artificial intelligence is quietly rewriting the way people watch sports. Real-time AI-generated highlights, automated intelligent commentary, personalized content recommendations, and immersive multi-angle viewing modes are now standard features on many streaming platforms, particularly in China’s fast-evolving digital sports ecosystem. Yet despite the rapid proliferation of these technologies, surprisingly little empirical evidence has existed on whether audiences actually accept them — and whether the technology genuinely improves the viewing experience. A new study published in Frontiers in Sports and Active Living provides some of the most detailed answers to date, using a classic psychological framework to map exactly how viewers’ perceptions of AI translate into loyalty and satisfaction.

The research, conducted by Liuxin Li, an independent researcher based in Guangzhou, China, applied the Technology Acceptance Model, or TAM, to AI-based sports broadcasting services. TAM, first formally proposed by Fred Davis in 1989 and rooted in the earlier Theory of Reasoned Action, holds that people’s acceptance of new technology hinges on two core beliefs: perceived usefulness, the sense that a system improves performance or experience, and perceived ease of use, the sense that operating it requires minimal time and cognitive effort. These beliefs shape a user’s overall attitude toward the technology, which in turn predicts behavioral intention — the willingness to keep using, recommend, and pay for the service.

What makes the new study distinctive is its addition of a second outcome variable alongside behavioral intention. The author introduced perceived viewing-experience enhancement, a construct capturing audiences’ subjective evaluations of improved match comprehension, emotional immersion, interactivity, and overall viewing satisfaction. The reasoning is that sports broadcasting is fundamentally a hedonic, experience-centered form of media consumption, unlike utilitarian technologies such as office software where task completion is the primary goal. In entertainment contexts, the core value audiences seek is the enhancement of the experience itself, so measuring only adoption intention risks overlooking what viewers actually gain.

To test the model, the study collected anonymous questionnaire data from March to April 2026 through online media platforms, using convenience and snowball sampling. Respondents first had to pass a mandatory screening question confirming they had hands-on experience with AI-enhanced sports broadcasting functions, filtering out participants with only superficial exposure. From an initial pool of 792 submitted questionnaires, rigorous multi-stage cleaning removed 41 duplicate submissions, 95 careless or logically invalid responses, and 26 responses from participants under 18 years old, leaving a final analytic sample of 630 valid adult respondents.

The sample skewed young and educated: 62.6 percent were male, 52.5 percent were aged 18 to 24, and 63.1 percent held bachelor’s degrees. All constructs were measured on five-point Likert scales adapted from well-established technology acceptance and media effects instruments, translated into Chinese and back-translated to ensure conceptual equivalence. Structural equation modeling, implemented in the R statistical environment using the lavaan package, was then employed to test the hypothesized relationships among the five latent variables.

The psychometric results were strong. Cronbach’s alpha coefficients exceeded 0.7 for every construct, with an overall scale reliability of 0.921. Confirmatory factor analysis showed acceptable model fit, with a comparative fit index of 0.942 and a root mean square error of approximation of 0.046. Convergent and discriminant validity were confirmed through average variance extracted values above 0.50, the Fornell-Larcker criterion, and heterotrait-monotrait ratios all below the strict 0.85 cutoff. Common method bias, a frequent concern in self-report surveys, was assessed with Harman’s single factor test and a latent common method factor technique, both indicating it was not a serious threat to the findings.

The structural results supported every hypothesis. Perceived ease of use positively predicted perceived usefulness with a standardized coefficient of 0.623, confirming that smoother, more intuitive interfaces make AI features seem more valuable. Both perceived ease of use and perceived usefulness significantly boosted usage attitude, and usage attitude in turn strongly predicted both behavioral intention and perceived viewing-experience enhancement, with coefficients of 0.689 and 0.703 respectively. Notably, usage attitude played a full mediating role: the perceptual antecedents influenced the two outcome variables entirely indirectly, with no significant direct paths, a specification confirmed by nested model comparisons using chi-square difference tests and information criteria.

Bootstrap mediation tests with 5,000 resamples and bias-corrected accelerated 95 percent confidence intervals reinforced these conclusions. The indirect effect of perceived ease of use on behavioral intention through usage attitude was significant, as was the corresponding pathway to viewing-experience enhancement. Chained sequential indirect effects, in which ease of use first builds perceived usefulness and then attitude, were also significant. Among the demographic controls — gender, age, education, and weekly sports-viewing duration — only viewing duration was positively associated with behavioral intention, and robustness analyses without the controls produced nearly identical results.

The theoretical implications are twofold. First, the study provides context-specific empirical validation of TAM in a non-Western, entertainment-oriented digital sports environment, extending a framework whose explanatory power has been documented across more than 2,300 publications in the past decade. Second, by treating perceived viewing-experience enhancement as a parallel endogenous outcome rather than a mere afterthought, the research moves TAM beyond predicting whether audiences intend to use a technology, toward explaining how technology-related perceptions translate into subjectively experienced gains in comprehension, immersion, and satisfaction. The author cautions, however, that the cross-sectional design cannot establish causal ordering, the self-report measures carry inherent common-method risks, and the young convenience sample limits generalizability to older demographic subgroups.

For broadcasters, media platforms, and AI service developers, the practical message is clear. Intuitive interfaces and low-complexity interaction designs remain foundational, because perceived ease of use is a precondition for everything downstream. Features must deliver tangible, observable viewing benefits — supporting match comprehension, personalizing content, and supplying relevant event information — since sophisticated AI that yields no perceptible gain will not cultivate positive attitudes. And because usage attitude is the psychological bridge to both continued engagement and enhanced experience, product iterations should focus on how functional design shapes users’ overall evaluative stance toward AI-driven viewing, not merely on raw technical performance. As AI continues to permeate live sports coverage worldwide, the study suggests that winning audiences depends less on the sophistication of the algorithms and more on whether viewers feel the technology genuinely makes watching the game better.

Subject of Research: Technology Acceptance Model applied to audience acceptance of AI-based sports broadcasting services

Article Title: Perceived viewing-experience enhancement and audience acceptance: applying TAM to AI-based sports broadcasting services

Article References: Perceived viewing-experience enhancement and audience acceptance: applying TAM to AI-based sports broadcasting services. (n.d.). https://doi.org/10.3389/fspor.2026.1945033

Image Credits: AI Generated

DOI: 10.3389/fspor.2026.1945033

Keywords: artificial intelligence, sports broadcasting, technology acceptance model, behavioral intention, viewing experience, structural equation modeling, audience acceptance, sports media, usage attitude, China, perceived usefulness, perceived ease of use