Consumer preferences for artificial intelligence-enhanced products: Differences across consumer segments, product types, and countries

Products with artificial intelligence (AI) functions perform thinking, decision-making, and social tasks, similar to human assistants. Drawing on this similarity to complement established theories of technology acceptance, this research develops hypotheses about consumers' preferences to buy or avoid AI products and the variation of these preferences across consumers and AI product types, which is particularly novel in the literature. Three studies from China, Japan, and the U.S. find that AI-specific privacy gain (especially for unhealthy consumers and AI product owners) and social connection (especially for lonely and extraverted consumers) have positive effects on intentions to purchase AI products. Moreover, AI-specific privacy concerns (especially for consumers with a high need for cognition and for products designed to resemble a living organism) and AI-specific violence concerns (especially for less open and less lonely consumers and for products designed as objects) negatively influence intentions to purchase AI products. Comparing the results across countries, AI privacy gain has the strongest motivational effect in the U.S., an individualistic country, and AI privacy/violence concerns have the strongest deterrent effect in Japan, a risk-averse country. Furthermore, AI social connection has the weakest effect in China, a collectivist country that may offer more conventional, non-AI social connection opportunities. • This research explains consumer preferences for AI-enhanced products. • Expected AI privacy gain attracts consumers (stronger for unhealthy consumers). • AI social connection attracts consumers (stronger for lonely/extraverted consumers). • AI privacy (stronger for need for cognition) and violence concerns cause aversion. • An AI product design as an organism vs. an object alters the impact of AI concerns.