Testing a Perceptual Factors Model of Young Users’ Perceived Emotional Recovery from Generative-AI-Assisted Healing Short Videos

Abstract

With the integration of generative artificial intelligence into short-video production, generative-AI-assisted healing short videos have become an emerging form of media content used by young audiences for everyday emotional regulation.However, the perceived emotional recovery associated with such videos may depend not only on technological novelty, but also on users' sensory, emotional, and immersive experiences.This study developed and tested a perceptual factors model to identify the determinants of young users' perceived emotional recovery from generative-AI-assisted healing short videos.A sequential mixed-methods design was employed.In the qualitative phase, semi-structured interviews were conducted with 15 young short-video users aged 18-35 who had prior exposure to generative-AI-assisted healing content.The interview transcripts were analyzed through sentence-by-sentence open coding, comparison and consolidation of semantically similar codes, and thematic categorization.This process identified five perceptual factors: sensory soothing, content empathy, AI aesthetic novelty, perceived authenticity, and viewing immersion.In the quantitative phase, 420 questionnaires were distributed, 386 were returned, and 358 valid responses were retained for analysis.Structural equation modeling showed that sensory soothing (β = 0.214, p = 0.003), content empathy (β = 0.286, p < 0.001), AI aesthetic novelty (β = 0.137, p = 0.028), perceived authenticity (β = 0.195, p = 0.006), and viewing immersion (β = 0.331, p < 0.001) all had significant positive effects on perceived emotional recovery.Viewing immersion exerted the strongest influence, followed by content empathy.Together, the five factors explained 61.2% of the variance in perceived emotional recovery.These findings indicate that the restorative value of generative-AI-assisted healing short videos depends less on visual novelty alone than on their ability to create an immersive, emotionally responsive, credible, and low-stimulation viewing environment.The study clarifies the perceptual mechanisms underlying emotional communication in generative-AI-assisted media and offers practical guidance for designing healing-oriented short-video content.