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

Pei Jialiang; Liu Shanshi; Zhong Chuyan; Chen Yifan

doi:doi:10.16538/j.cnki.fem.20210818.103

Summary

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.