AI vs. Human Streamers on Consumer Purchase Intention: An Asymmetric Mechanism Model with Two Boundary Conditions among Malaysian Consumers

Abstract

Simple comparisons of the effects of AI streamers and human streamers in live-streaming e-commerce have been widely examined, yet the differentiated psychological mechanisms underlying their effects remain unclear. This study proposes an asymmetric mechanism model: human streamers primarily form an affective advantage through perceived authenticity and social presence, whereas AI streamers form functional compensation through perceived professionalism. Product type and AI acceptance are incorporated as boundary conditions at the contextual and individual levels, respectively. A two-stage research design was adopted. Study 1, based on a survey of Malaysian consumers (N = 372), examined asymmetric mediation and preliminary robustness across ethnic backgrounds. Study 2, using a 2 × 2 online scenario experiment (N = 314), validated the causal effects. The results show that the advantage of human streamers mainly stemmed from the authenticity mechanism, which was the most stable mechanism across both studies. The social presence mechanism appeared only to a limited extent in the context of hedonic products. The professionalism-based compensation mechanism of AI streamers was significant mainly in the context of utilitarian products. AI acceptance primarily moderated the authenticity mechanism, indicating that consumers with higher AI acceptance were less sensitive to AI streamers’ authenticity disadvantage. Its moderating effect on the professionalism mechanism was only marginally significant, suggesting that the role of AI acceptance in shaping professionalism evaluations requires further examination. This study shifts the research focus from “which type of streamer is more effective” to “through which mechanisms streamer effects occur,” thereby providing implications for human–AI collaborative streamer strategies in Southeast Asia.