Understanding Fairness and Favouritism in AI-Based HR Appraisals: A Mixed-Methods Analysis of Employees in the Indian IT Sector

Abstract AI-driven HR systems are increasingly adopted across organisations, yet concerns regarding transparency, fairness, and hidden bias continue to affect employee experiences. This mixed-method study examines how algorithmic transparency influences perceived fairness and workplace favouritism, and how digital literacy moderates these relationships within the Indian IT environments. Qualitative insights from 20 phenomenological interviews further revealed themes of black-box opacity, data visibility bias and algorithmic favouritism, highlighting how contextual sensitivity shapes negative user experiences. Quantitative data from 400 employees were analysed using PLS-SEM, revealing that transparency significantly enhances fairness perceptions, which in turn reduces perceived favouritism, confirming fairness as a key psychological mediator. Digital literacy emerged as a strong moderator, indicating that employees with higher literacy interpret AI outputs more accurately, reducing suspicion and algorithmic anxiety. Triangulating across methods, the study concludes that fairness is not achieved through automation alone. Rather, it arises from transparency, interpretability, and digital empowerment. Findings provide actionable implications for AI governance, HR design, and organisational policy, underscoring the importance of human-centred algorithmic management.