Should manufacturers adopt AI-generated livestreamers to replace human livestreamers? Impact of consumer disappointment aversion

Purpose This study focuses on the manufacturer's choice between human streamers and AI-generated streamers in an online-offline dual-channel supply chain under consumer disappointment aversion. Design/methodology/approach To address whether the adoption of AI-generated livestreamer is better for the manufacturer, we consider two scenarios: Human livestreamer sales (Scenario H) and AI-generated livestreamer sales (Scenario A). Consumer disappointment aversion is embedded in the utility specification, enabling us to derive demand and formulate a Stackelberg pricing game between the manufacturer and the retailer. A comparison of the equilibrium results across the two scenarios reveals how each supply chain member values human versus AI-generated livestreamers. Findings The manufacturer replaces the human livestreamer with the AI-generated one under conditions of (1) moderate consumer acceptance of AI-generated livestreamers and low disappointment aversion level or (2) high consumer acceptance of AI-generated livestreamers. Moreover, interestingly, the increase in the commission rate for the human livestreamer does not always result in the manufacturer preferring the AI-generated livestreamer, but rather impacts the manufacturer's preferences non-monotonically. Finally, six extensions indicate that the findings regarding the manufacturer's livestreamer selection remain robust under certain conditions. Originality/value This study is the first to integrate AI technology, consumer disappointment aversion, and a dual-channel structure into a unified framework to investigate supply chain pricing and livestreamer selection decisions. The findings offer theoretical guidance for dual-channel supply chain members in livestream commerce regarding product pricing, livestreamer choice, and managing consumer disappointment aversion behavior.