AI Application and Privacy Concerns: Comparing Users' Perceptions of AI Chatbots and Human Agents

The rapid development of artificial intelligence (AI) has attracted wide attention and generated heated discussions about user privacy. As an AI application, a chatbot is a conversational agent with natural language processing ability to imitate human conversation. Since chatbots cannot replace humans in the performance of more complicated or personalized tasks, human agents still co-exist with chatbots in the business or other sectors. A chatbot can be embedded in different platforms or through different channels to implement particular tasks. It can also develop a long-term relationship with and even become an emotional companion to its users. In expanding the existing scholarship regarding human-robot interaction, this study will cross-culturally explore users' privacy concerns with regard to two types of chatbots, namely, task-oriented and companion-oriented chatbots and their different impacts on people's usage intentions. It will also examine users' perceived trust in the expertise, reputation, or technology of chatbots and their intentions in using chatbots and human agents.Studies have shown that cultural or regional differences regarding users' privacy concerns should not be ignored when conversational agents are involved. People in Mainland China are generally more receptive to new technologies. They are more willing to provide their personal data in exchange for the services and products they need. However, people in Hong Kong and the U.S. tend to be more cautious when using new smart devices or applications. Therefore, the focus of this study will be on Hong Kong, Mainland China, and the U.S. In the first year, in-depth interviews with chatbot designers and relevant enterprises will be conducted. Meanwhile, panel group studies that focus on the users will be conducted in Hong Kong and Mainland China in order to prepare for the online experiment. In the second year, the online experiment will be conducted in the three regions based on having 2 (chatbot vs. human) x 2 (task-oriented vs. companion-oriented) conditions to understand users' perceived trust, privacy concerns and usage intention. After the online experiment, the panel group studies will be conducted again in order to supplement the above quantitative findings. This study will contribute to scholars' and practitioners' understanding of chatbots in cross-cultural digital communication environments by investigating users' perceptions of the communication entity (chatbot) and their usage intention when interacting with them. In delineating the notions of perceived trust and privacy concerns, this study aims to extend research in the field of human-machine communication along with the societal implications.