Consumer Responses to Ethical and Privacy Concerns in AI Chatbots: Evidence From Offensive Language, Gender Bias, and Digital Privacy

Due to the rapid innovation of artificial intelligence (AI) and big data, AI chatbots have become increasingly influential computational social systems (CSS) in everyday life. However, the expansion of interactions between AI chatbots and consumers has raised significant ethical and privacy concerns. Using data from an online survey, this study empirically examines how consumer concerns about AI ethics and privacy influence their intention to use AI chatbots. These concerns are categorized into three types: 1) offensive language (generated by AI chatbots or users); 2) gender-biased language (from either source); and 3) digital privacy violations. The findings suggest that prior experience with AI chatbots is the most significant positive influence on respondents’ intention to use them. In contrast, all concern variables negatively influence respondents’ intention to use AI chatbots. Moreover, this intention varies depending on the type and source of unethical behavior. Specifically, concerns about user-generated offensive language directed at AI chatbots have a stronger negative effect than concerns about offensive language from AI directed at users. Conversely, concerns about AI-generated gender-biased language toward users have a stronger negative impact than concerns about such language from users toward AI. Finally, respondents express greater sensitivity to AI chatbots’ violations of their own digital privacy than to violations affecting other users.