Resume-ing Control: (Mis)Perceptions of Agency Around GenAI Use in Recruiting Workflows

Problem
This preprint addresses the gap in empirical understanding of how generative AI (genAI) systems impact the perceived agency and control of professionals in high-stakes decision-making contexts, specifically within recruiting workflows. While existing literature emphasizes the role of genAI as an assistive tool, there is limited exploration of how its integration affects the autonomy and decision-making processes of recruiters. The study aims to elucidate the nuanced dynamics between human agency and AI influence in recruitment, an area that has not been thoroughly investigated.

Method
The authors conducted qualitative interviews with 22 recruiting professionals to gather insights into their experiences and perceptions regarding the use of genAI in their workflows. The study employed a grounded theory approach to analyze the data, allowing for the emergence of themes related to control, agency, and the perceived impact of genAI on recruitment processes. The focus was on understanding how genAI systems shape decision-making, from job definition to candidate evaluation, and the extent to which recruiters feel they retain authority over these processes.

Results
The findings reveal a significant disconnect between recruiters’ perceptions of their authority and the actual influence of genAI on their workflows. While participants believed they maintained final decision-making power, they reported that genAI systems subtly dictated the foundational elements of recruitment, such as job descriptions and evaluation criteria. Despite the integration of genAI, efficiency gains were reported as marginal, with many recruiters expressing concerns about deskilling and a loss of meaningful oversight in hiring decisions. The study highlights that the decision to adopt genAI was often driven by external pressures, including directives from management and the need to enhance productivity in response to applicant use of AI tools.

Limitations
The authors acknowledge several limitations, including the small sample size and the qualitative nature of the study, which may limit the generalizability of the findings. Additionally, the research does not quantify the specific efficiency gains or deskilling effects, relying instead on subjective reports from participants. The study also does not explore the long-term implications of genAI adoption on recruitment practices or the potential for bias introduced by AI systems.

Why it matters
This research has significant implications for the responsible integration of genAI in hiring processes. It underscores the need for transparency in how AI systems influence decision-making and the importance of maintaining human oversight to prevent deskilling among recruiters. The findings suggest that organizations must critically evaluate the adoption of genAI tools, considering not only productivity gains but also the potential erosion of agency and expertise among human decision-makers. This work lays the groundwork for future studies on the ethical use of AI in recruitment and the development of frameworks that prioritize human agency in AI-assisted decision-making.

Authors: Sajel Surati, Rosanna Bellini, Emily Black
Source: arXiv:2604.26851
URL: https://arxiv.org/abs/2604.26851v1