Abstract:
This paper explores the unrealized potential of artificial intelligence in human resource management and suggests how progress might be made.
The authors identify several key challenges with using data science techniques to enhance HR practices. These include the complexity of HR phenomena, the constraints imposed by small data sets, the ethical questions associated with fairness and legal considerations, and the employee reaction to management through data-based algorithms.
The authors then propose some practical responses to these challenges—specifically, causal reasoning, randomization, and process formalization—that they argue are both economically efficient and socially appropriate ways to integrate data analytics into the management of employees.
Key Points:
- AI is increasingly applicable to HR decisions, but there are some constraints and concerns that still need to be addressed.
- Algorithmic HR decisions are less biased and more accurate than human ones but are often perceived as less fair because the biases they do have are inherently systematic and easily understood.
- Key ways to increase fairness and acceptance of AI in HR: go beyond association to establish causal relationships between traits and outcomes, introducing randomization, and formalize processes.
Takeaway:
AI-augmented decision making is the future of human resource management, but it’s implementation needs to appropriately address the inherent tension that exists between efficiency and fairness.
