May AI help you? A quantitative study investigating a potential means to overcome the disclosure paradox of chatbots in the service encounter

Sammanfattning: The use of artificial intelligence solutions such as chatbots in service encounters is gaining in popularity due to the potential efficiency gains from automating routine interactions. However, the desired benefits are dependent on the identity of the virtual service agent not being disclosed, as customers who know that they are interacting with a chatbot perceive it to be less competent and trustworthy, eliminating all efficiency gains. Yet the choice to omit disclosure is not only deceitful, but also increasingly restricted by law, leading to a disclosure paradox. A significant research gap exists since little attention has been paid to identity disclosure of chatbots. Since mere identity disclosure leads to negative attitudes, the purpose of this study is to explore whether a variant of the second-order disclosure can affect humans' resulting attitudes as compared to first-order disclosure. The proposed second-order disclosure includes the chatbot identity and a competence signal. A quantitative study with 1000 participants was conducted through an experiment that divided the treatment and control group based on whether the disclosure of the chatbot identity included a competence signal or not. The study found that the competence signal was unsuccessful in impacting attitudes. However, the importance of perceived competence and trust for attitude towards the chatbot and the brand was determined. The study found that perceived chatbot competence leads to a higher attitude towards the chatbot, through the mediating effect of trust, with spillover effects on brand liking and the intention to use the chatbot. The findings indicate the challenge and importance of further investigating means to overcome the disclosure paradox. Therefore, this study highlights theoretical implications of the practical dilemma.