Bridging the trust gap: anthropomorphism, chatbot features, and system features in AI-chatbot adoption across cultures

Abstract

Abstract

As artificial intelligence (AI)-chatbots mimic human-like communication, they are able to reshape customer-business interactions. Due to a lack of user trust, their adoption falls below expectations. This study encompasses the dual-pathway model of trust and discovers how human-like features (empathetic response, intelligence, social presence, interactivity) impact users’ trust via anthropomorphic perceptions (affective route) and how system features (usefulness, service quality, ease of use) contribute to trust via the cognitive route. Overall, forty English-language publications reporting correlation coefficients (or convertible statistics) and sample sizes published till December 2024 were identified from Scopus, EBSCO, and ProQuest. Their meta-analysis confirms the positive effects of system features on trust and chatbot features on anthropomorphism, which further enhances trust. Prior studies have not addressed cultural heterogeneity in AI-chatbot research due to limited geographical scope. This study introduces Eastern/Western cultures, individualism, and uncertainty avoidance as the moderating variables. Results reveal the stronger effect of perceived usefulness on trust and social presence on anthropomorphism in Western cultures. Also, anthropomorphism has a greater influence on trust in Eastern cultures. The positive relationship between anthropomorphism and trust is weakened as individualism increases. Also, as uncertainty avoidance increases, the positive relationship between perceived ease of use and trust is strengthened.