Abstract
As artificial intelligence (AI) reshapes the workplace, its impact on diversity, equity, and inclusion (DEI) has been examined primarily through the lens of what AI actually does (i.e., system design, outputs). This review shifts the focus to humans' subjective beliefs about AI, identifying mechanisms through which three dimensions of AI-beliefs (competence, warmth, and morality) may affect DEI goals: improving representation of underrepresented groups, closing career gaps across groups, and fostering inclusion. Our synthesis shows that, independently of what AI does, AI-beliefs can disadvantage underrepresented groups in some contexts, while creating opportunities for bridging human subgroups in others. This suggests that achieving DEI goals in AI-integrated workplaces requires attention not only to the properties and behavior of AI, but also to employees' beliefs about AI itself.