AI IN RECRUITMENT A Study on Readiness and Responsibility in the Face of Regulation

Abstract

This study examines how artificial intelligence (AI) is used in recruitment in Sweden, with a particular emphasis on regulatory compliance, technical complexity, responsibility and understanding. Based on 10 semi-structured interviews with recruiters, the study reveals that AI is most commonly used in screening, generating recommendations, documentation and drafting job advertisements. However, the results show that the screening phase and AI-generated recommendations are subject to the highest regulatory scrutiny, requiring greater human oversight. While human involvement is emphasized in a way that complies with regulatory frameworks such as the AI Act and the GDPR, fully automated AI decisions still occur, specifically in the screening phase. In addition, the findings suggest that what is perceived as human oversight may be superficial during AI-generated recommendations. Trust in AI appears stronger when the output aligns with the recruiter ´s own judgment. This directly conflicts with the AI Act and GDPR, restricting the use of AI in high-risk domains, such as recruitment. Furthermore, the results reveal a lag in organizational readiness through the lack of structured education, restricting recruiters' understanding of the underlying logic of AI systems. These findings reveal a risk that organizations adopt symbolic forms of responsibility, namely accountability and explainability but also human oversight. Consequently, the fairness of recruitment processes risks being questioned. Therefore, this study offers advice for practitioners to ensure a responsible use of AI in recruitment. The advice is: Ensure a deep understanding of AI output logic, invest time and resources and involve lawyers early in the implementation process.