Not all AI Chatbots are Created Equal: The Case of a Soft Skills Coaching Bot in Education

Chatbots are becoming a daily companion of human life. With the recent advances in generative AI, it is easier than ever to design and develop chatbots. However, we argue that not all AI chatbots are created equal. Thus, in this, one of a kind, empirical study we investigate the factors that impact users' intention to use three versions of a soft skill chatbot. We also offer a comparison of them to understand if the generative AI enabled chatbots lead to more intention to use than traditional AI chatbots. To address our research questions, we extended the Unified Theory of Acceptance and Use of Technology model, by incorporating Perceived Risk and Perceived Trust and by adding the generative AI-enabled chatbots as treatments in the model. Our findings reveal noteworthy disparities between the three versions of chatbot, offering both theoretical and practical contributions. Among the theoretical contributions we found that Effort Expectancy has negative impact on intention to use for generative AI chatbots, which is challenging extant literature. Moreover, we identified key mediators (e.g., Perceived Trust, Performance Expectancy and Effort Expectancy), shedding light on critical factors influencing users' intention to engage with these technological tools. Among the practical contributions we offer actionable insights for educators and practitioners seeking to integrate AI chatbots into educational settings.