Reconceptualizing the Automation–Augmentation Tension in AI-Enabled Talent Selection: A Stage-Embedded Theory

Artificial intelligence (AI) is increasingly embedded in talent selection, yet its deployment remains characterized by a persistent tension between automation and augmentation. While existing research often treats this choice as an organizational-level strategy or a temporal progression, it frequently overlooks how institutional sensitivities and task characteristics vary across the recruitment lifecycle. This research examines how AI is configured as automation, augmentation, or hybrid collaboration across specific recruitment stages and why organizations repeatedly transition between these configurations. Adopting a comparative, multi-case theory-building approach, the paper analyzes documented AI deployments across standardized selection stages, including applicant screening, assessment, and final selection. The findings identify three recurring stage-embedded mechanisms. Early recruitment stages stabilize efficiency-driven automation under conditions of high task analyzability and low immediate accountability. Intermediate stages intensify paradoxical tensions, producing unstable hybrid configurations as organizations attempt to balance scalability, validity, and fairness. Late stages institutionalize accountability-driven augmentation, structurally constraining AI autonomy due to normative and legal requirements for human responsibility, regardless of technical performance. Across these stages, organizations engage in "re-augmentation" as a legitimacy repair mechanism to restore trust and institutional alignment following perceived algorithmic failures or ethical concerns. By reconceptualizing the automation–augmentation paradox as a stage-embedded phenomenon, this work demonstrates that contradictory logics are distributed spatially across the hiring process. These insights refine task complementarity theory by establishing accountability as a fundamental boundary condition that limits human–AI collaboration in high-stakes human resource decisions.