Generative artificial intelligence (AI) is increasingly used to produce multimedia content, raising important questions about how users evaluate and respond to such content. While prior research has examined AI transparency and technology adoption, the cognitive mechanisms linking external cues to user trust and behavioral outcomes remain insufficiently understood. Grounded in the heuristic–systematic model (HSM), this study investigates how transparency labels and presentation quality influence cognitive processing, perceived authenticity, trust, and perceived risk in AI-generated video content. A 2 × 2 between-subjects experiment was conducted with 617 participants. The results show that AI transparency labels reduce perceived authenticity but do not directly affect trust or adoption intention. Higher presentation quality reduces reliance on heuristic processing. Perceived authenticity is associated with lower cognitive load and greater systematic processing, systematic processing is positively associated with trust, whereas heuristic processing is negatively associated with trust. Trust is positively associated with both adoption intention and sharing intention, whereas perceived risk is negatively associated with these outcomes. These findings highlight the central roles of cognitive processing and trust–risk dynamics in shaping user responses to AI-generated content and provide implications for transparency design and trust development in AI-mediated environments.
Liu et al. (Wed,) studied this question.