Ai Anxiety as a Positive Moderator in Banking Chatbot Continuance Usage Intention: Pls-Sem and Ann Analysis

Abstract

Despite the widespread application of artificial intelligence (AI) in China's banking sector, user recognition of banking e-service chatbots remains low. The purpose of this study is to validate an extended model of banking e-service chatbots continuance usage. The model is constructed by extending the information system (IS) success model and the expectation confirmation model of information system continuance (ECM-ISC), incorporating three additional constructs: trust, perceived risk, and AI anxiety. A two-stage analysis using partial least squares structural equation modeling with artificial neural networks is conducted with data from 331 users in China. The findings suggest that user satisfaction is positively affected by the design dimensions of banking e-service chatbots, but trust is not impacted by service quality. Continuance usage intention is influenced by satisfaction, trust, and perceived usefulness, and perceived risk has no significant effect on it. The findings show that AI anxiety positively moderates the influence of satisfaction as well as trust on the continuance of usage intention. Additionally, based on self-congruity theory, multigroup analysis is conducted to explore differences between cross-platform and single-platform users. By extending and validating the IS success model and ECM-ISC, this study explores the factors that influence the intention to continue using banking e-service chatbots and provides valuable insights into their design.

Keywords: Banking e-service chatbot, Continuance usage intention, AI anxiety, PLS-SEM-ANN