Study Of AI Trust Model With Risk Level Of Introducing Scenarios

Driven by the fourth industrial revolution,various products led by artificial intelligence and machine learning have opened up new directions of innovation for economic growth,and human-computer interaction(HCI)as the cornerstone of product design has evolved into a more flexible machine autonomy model.In this context,trust has become a topic of increasing concern in HCI,and the lack of trust in HCI has become a key factor leading to the slow development of AI applications.In order to solve the problems of user trust in HCI,this study systematically composes the theories and related research results in HCIrelated fields,deeply understands the current situation,shortcomings and feasible directions of research,and proposes a unified risk perspective and incorporates the user’s desirability tendency into the model,and then establishes a comprehensive trust research model that includes user characteristics,AI system characteristics and scenario risk level.The comprehensive trust research model incorporating user characteristics,AI system characteristics,and scenario risk levels is then established,and has some relevance to the actual design of AI products.In order to validate the research model proposed in this study,mature AI application products in the market are used,and data are collected and validated through online intergroup experiments.The results show that the scenario risk level,user’s likability tendency,AI interpretability,machine performance,and anthropomorphism all have some influence on the trust of humancomputer interaction.Among these effects,scenario risk level and user’s desirability tendency have a moderating effect on trust.This implies that AI interpretability,anthropomorphism,and machine performance have different effects on the trust level of HCI in different scenarios.These f ndings are not only important for understanding the trust relationship in human-computer interaction,but also provide valuable guidance for developing more trustworthy AI systems in the future.In order to verify the feasibility of the research model,this study employs a data research method of within-group experiments to validate the model.The research in this paper not only expands the scope of research in the field of human-computer interaction,but also lays a solid foundation for further research to follow.The research results have implications for both academic research and practical applications in the field of human-computer interaction.In addition,this paper provides useful insights and guidance for developing more reliable and trustworthy AI systems,and also has a certain role in promoting the development and application of AI technology.