Affiliation:
1. Universitas Trisakti, Indonesia
Abstract
The use of Artificial Intelligence (AI) technology is often encountered in everyday life, especially in the world of business marketing, namely chatbots which are part of the implementation of Natural Language Processing (NLP). However, in its application, it is still felt that it cannot meet the needs of consumers for the specific questions they often ask, especially for those who use health teleconsultation services. In this study, the intelligence variable is added as an additional variable to answer the level of user trust in the chatbot. The data used in this study are quantitative with the target respondents being users of health teleconsultation services with a minimum user experience of 1 year. A total of 178 respondents met the criteria from a total of 238 respondents. In this research, three variables are developed. For each dependent variable, empathy, friendliness, and intelligence are factors moderated by task complexity and chatbot disclosure to determine the results of the independent variable, which is trust towards the chatbot. Based on the results of the outer model and inner model analysis using SmartPLS, it can be concluded that the addition of intelligence variables has a positive effect on user trust in chatbots. In addition, the level of chatbot complexity is also able to mediate the relationship between intelligence and user trust in chatbots. However, chatbot disclosure has a negative effect as mediating the relationship between intelligence and user trust in chatbots. In the application of using chatbots, the level of intelligence may have an effect on user trust, but natural human attitudes such as friendliness from chatbots do not affect user trust.