This study examines the discursive construction of algorithms and its role in labor management in Chinese live-streaming industry by focusing on how intermediary organizations (Multi-Channel Networks, MCNs) actively construct, stabilize, and deploy particular interpretations of platform algorithms as instruments of labor management. Drawing on a nine-month ethnographic fieldwork and 44 interviews with live-streamers, former live-streamers, and MCN staff, we examine how MCNs produce and circulate structured interpretations of platform algorithms across organizational settings. We show that MCNs articulate two asymmetric yet interconnected forms of algorithmic interpretations. Internally, MCNs managers approach algorithms as volatile and uncertain systems and adopt probabilistic strategies to manage performance and risk. Externally, in interactions with streamers, MCNs circulate simplified and prescriptive algorithmic narratives that frame platform systems as transparent, fair, and responsive to individual effort. These organizationally produced algorithmic interpretations are embedded into training materials, live-streaming performance metrics, and everyday management practices. Through these mechanisms, streamers internalize responsibility for outcomes, intensify self-discipline, and increase investments in equipment, performing skills, and routines to maintain streamer-audience relationship, while accountability for unpredictable outcomes is increasingly shifted away from managers and platforms. This study contributes to CSCW and platform labor research by demonstrating how discursively constructed algorithmic knowledge can function as an intermediary infrastructure of soft control, shaping how platform labor is regulated, moralized, and governed in practice.

翻译：本研究聚焦于中介组织（多频道网络，MCNs）如何主动构建、稳定并部署对平台算法的特定解读，以之作为劳动管理工具，从而考察算法的话语建构及其在中国直播行业劳动管理中的作用。基于为期九个月的民族志田野调查以及对直播主播、前主播和MCN员工的44次访谈，我们考察了MCN如何在组织环境中生成并传播对平台算法的结构化解读。研究表明，MCN阐明了两种不对称却又相互关联的算法解读形式。在内部，MCN管理者将算法视为多变且不确定的系统，并采用概率性策略来管理绩效与风险。在外部，在与主播的互动中，MCN传播简化且规范性的算法叙事，将平台系统描绘为透明、公平且对个人努力有响应的。这些由组织生成的算法解读被嵌入培训材料、直播绩效指标和日常管理实践中。通过这些机制，主播内化了结果责任，加强了自我规训，并增加了在设备、表演技能和维持主播-观众关系的日常投入上的投资，而对不可预测结果的责任则日益从管理者和平台方转移。本研究通过展示话语建构的算法知识如何作为一种软控制的中介基础设施发挥作用，塑造平台劳动在实践中如何被规制、道德化和治理，从而为CSCW和平台劳动研究做出了贡献。