Neural Network-enhanced Analysis: How Risk Perceptions of Artificial Intelligence Diffuse
Han Wu, Tian Wei, Michael Dowling · Academy of Management Proceedings · 2023
Previous research has shown that a firm’s use of artificial intelligence (AI) in daily management tasks is influenced by their employees’ risk perceptions regarding AI implementation. These risk perceptions can be categorized as functional, anthropogenic, and financial. As employees’ risk perceptions are partly derived from their cognition of AI’s features, we proposed that there is a positive correlation between each risk perception category. In addition, we anticipated that employee work experience would also enhance the associations between these categories. Thus, we conducted two studies within the accounting industry using linear correlation and an artificial neural network. Our results indicate that employees with high levels of functional and anthropogenic risk perception also have higher levels of financial risk perception. Additionally, although working experience seemed irrelevant to the relationships between risk perceptions, it determines the employees’ perception of financial risks in a non-linear and complex manner. Our findings suggest that firms should improve their employees’ abilities to perceive risks to enable the efficient implementation of AI technologies within their firms.