ABSTRACT Discrimination is both an objective reality and a subjective perception. While firms increasingly use AI for decisions like hiring, little is known about how AI versus human decision‐makers affect perceived discrimination. Across seven studies ( N = 1757), AI decisions are consistently seen as less discriminatory than identical human decisions across various contexts of discrimination. Study 2 shows this is mediated by inferred intentions: AI is viewed as neutral, while humans are seen as selfish or malicious. Study 3 rules out perceived predictability as an alternative explanation. Studies 4 and 5 find that high‐adaptivity algorithms further reduce perceived discrimination, again mediated by intentions. This paradoxically favourable perception of AI decisions masks underlying biases, complicating efforts to detect and address discrimination. These findings highlight the critical gap between perceived and actual fairness in AI‐driven decisions, underscoring the need for transparency, rigorous auditing and effective governance mechanisms to align perceptions with objective equity.
Li et al. (Tue,) studied this question.