Evaluating metrics of fairness: a critical analysis of AI in hiring practices

AI-mediated hiring is increasingly recognized as a high-stakes domain under emerging global regulatory frameworks, prompting new questions about how notions of fairness and suitability are constructed and communicated. Drawing on research in algorithmic opacity and quantification, this study critically examines how AI hiring tools frame these concepts discursively. Through an analysis of technical documents, marketing materials, and public narratives, we examine how two major hiring platforms – LinkedIn Recruiter and HireVue – frame and communicate ideas of fairness and suitability. We identify three paradoxes central to AI-mediated hiring: the discrepancy between claimed algorithmic fairness and potential embedded biases, the tension between efficiency and human discretion, and the challenges of cultural inclusivity in globally marketed tools. We treat these materials as relational artifacts that mediate values, expectations, and accountabilities among platform vendors, employers, and job seekers. While these platforms aim to mitigate bias, they reveal deeper tensions around transparency, accountability, and the normative assumptions underpinning the evolving landscape of AI-mediated recruitment.