This study aims to identify the cognitive emotional mechanisms underlying why AI-based performance reproduction does not fully feel “authentic” to audiences, and to analyze the structure of audience authenticity perception. An AI-based performance reproduction video featuring Audrey Hepburn was used as a stimulus. The study does not seek to evaluate the authenticity of a specific traditional dance form, but focuses on how perceived visual artificiality influences authenticity loss and emotional distance. The research model draws on uncanny valley theory, perceptual realism, authenticity theory, and Walter Benjamin’s concept of aura. PLS-SEM analysis of 1,124 respondents revealed that unnatural gaze and facial expressions, violations of physical laws and detail errors, and excessively smooth skin representation all had significant effects on authenticity loss and emotional distance. Authenticity loss significantly affected emotional distance and partially mediated the relationship between visual artificiality and emotional distance. The findings suggest that audience experience in digital performing arts may be shaped more strongly by the perceived naturalness and authenticity of performance than by the level of visual reproduction itself.