This study offers a nuanced exploration into the intersection of expertise and AI-powered decision-making, particularly within the realm of high-volume recruitment. It leverages theory from the evolving discourse on relational expertise and human-AI interaction to examine how experts navigate, interpret, and sometimes challenge AI tool outputs. Through in-depth interviews with 42 recruitment experts, the study focuses on the concept of algorithmic folk theories—the interpretive frameworks through which experts engage with algorithmic recommendations. Central to the study's findings is the range of perceptions among experts toward AI technologies, viewed through the lens of expert-AI pairings. These perceptions oscillate between viewing AI as either a complementary ally or a challenging rival, significantly shaped by organizational contexts. Factors influencing these views include oversight levels, trust in AI outputs, and the prioritization of AI tools in decision-making processes. Findings also reveal instances of algoactivism, where experts actively resist or workaround AI outputs to align with their professional judgment. In turn, algorithmic folk theories are interpretive frameworks informed by and situated within organizational structures.

中文翻译：

---

专家与人工智能配对：专家干预人工智能驱动的决策


这项研究对专业知识和人工智能驱动的决策的交叉点进行了细致入微的探索，特别是在大批量招聘领域。它利用有关关系专业知识和人机交互的不断发展的论述中的理论来研究专家如何导航、解释甚至有时挑战人工智能工具的输出。通过对 42 名招聘专家的深入访谈，该研究重点关注算法民间理论的概念，即专家参与算法推荐的解释框架。该研究结果的核心是从专家与人工智能配对的角度来看，专家对人工智能技术的看法范围。这些看法在将人工智能视为互补盟友或具有挑战性的竞争对手之间摇摆不定，很大程度上受到组织环境的影响。影响这些观点的因素包括监督水平、对人工智能输出的信任以及决策过程中人工智能工具的优先顺序。调查结果还揭示了算法激进主义的实例，即专家积极抵制或解决人工智能输出以符合他们的专业判断。反过来，算法民间理论是由组织结构提供信息并位于其内部的解释框架。