Delegating decision authority to AI in recruitment: an empirical study of trust, control and algorithmic accountability

Abstract

The widespread adoption of artificial intelligence (AI) in human resource (HR) practices has led to a growing number of organizations trying to understand the use of AI in recruitment. While AI is increasingly embedded in HR practices, few studies focus on how trust in AI is formed and how it shapes the adoption of AI in recruitment. This study employed a qualitative research design based on 21 semi-structured interviews with HR professionals working in large, internationally operating German family-owned organizations that use AI in recruitment. The data was analysed using a theory-guided qualitative approach following the Gioia methodology to construct a systematic data structure linking participants’ experiences to established theoretical dimensions of trust. The findings reveal that trust in AI is shaped by three interrelated dimensions: ability, integrity and benevolence. The study underscores how trust serves as a dynamic slider that governs the degree of decision authority delegated to AI systems, varying from automated screening for habitual tasks to human oversight for complex, cognitive tasks. Key tensions also emerge around power, control and trust breakdown. First, the study emphasizes that organizations should recognize that successful AI adoption depends not only on technical performance, but also on deliberate trust-building strategies. By investing in transparency mechanisms, such as explainable AI interfaces, organizations can get a clear understanding of how recruitment algorithms operate. Second, organizations should invest in structured training programs to develop the efficiency and confidence of HR professionals working with AI systems. This can be done by offering training that focusses on interpreting algorithmic outputs, recognizing potential biases and maintaining ethical oversight. Third, organizations should establish formal governance frameworks that define accountability, certification standards and internal usage guidelines, which can reduce uncertainty and promote consistent, responsible AI practices in recruitment practices. The study contributes to the technology acceptance model (TAM) and the algorithmic accountability theory by demonstrating how trust conditions the adoption of AI in recruitment. It theorizes how micro-level trust judgements are structurally conditioned by macro-level institutional safeguards and provides empirical insights into how HR professionals balance human judgement and algorithmic decision-making.