Can AI Algorithmic Decision-making Improve Employees’ Perception of Procedural Fairness?

With the development of digital technologies such as big data, artificial intelligence and machine learning, many organizations try to introduce AI algorithms into the decision-making process to avoid the subjective inherent biases. Although AI algorithms improve the scientificity of decision-making, they also raise the issue of fairness in the decision-making process. Considering that employees are the direct subject of HR decision-making, how employees view and evaluate the fairness of AI algorithmic decision-making process is very important. However, researches on the above questions are rather limited.Based on the procedural fairness theory, this study discusses the impact of different decision-making subjects (AI algorithms vs. supervisors) on employees’ perception of fairness in the HR decision-making process. Through two experiments, we find that: (1) The decision-making of AI algorithms makes employees have a lower perception of procedural fairness than that of supervisors. (2) Information transparency mediates the relationship between AI algorithmic decision-making and employees’ perception of procedural fairness. Employees believe that AI algorithmic decision-making has lower information transparency, which in turn produces a lower perception of procedural fairness. (3) Inclusive climate moderates the relationship between AI algorithmic decision-making and employees’ perception of procedural fairness. When employees perceive a high level of organizational inclusive climate, the effect of AI algorithmic decision-making on the perception of program fairness will be weakened.Overall, this study responds to the call for more research that bridges AI algorithmic decision-making and individual perception of procedural fairness by examining information transparency as a mediator and inclusive climate as a moderator in this relationship. The findings not only challenge the existing mainstream view that AI algorithmic decision-making does not exist a lower bias than human decision-making, but also offer useful advice to managers and provide implications for organizational HR decision-making practices in the digital and intelligent era.