当前学界普遍认为算法对新闻生产有重要的形塑作用，同时将其置于人与技术二元 对立的框架中讨论，发展出一套“算法神话”的话语。本研究则将算法视为人类与世 界之间关系的中介者，提出用“算法调节”的视角观察新闻透明性原则在算法新闻分 发平台的实践中所面临的障碍。研究通过田野观察和深度访谈发现，平台的组织架构和 物理空间带来了天然的挑战；人和算法作为“人-技混合体”的复杂性构成实际的“黑 箱”；使用者在算法的影响下也逐渐被“数据化”。基于上述情况，本研究沿着“技术 道德化”思路提出改进路径，包括把新闻透明性纳入算法设计的“常规”、加强算法利 益相关方的协作，以及用对话而非对抗的方式重新调整人和算法的关系等。

At present, the academic community generally believes that algorithms have a critical “informing” effect on news production, meanwhile, algorithms are discussed inside the ontological framework of the binary opposition between humans and technology, which shows a contradictory and dual “algorithmic myth”. This thesis is in favor of the concept of technology as the intermediary between human beings and the world, and proposes to switch from an “algorithmic myth” to an “algorithmic mediation” to observe whether news transparency is facing obstacles in the feld of algorithm-led news distribution. Through observation and indepth interviews, this paper fnds that the organizational structure and spatial, physical barriers of the platform bring challenges to the grounding of news transparency. Human and algorithm, as the cognition and practice of “human-technology associations”, constitute the actual “black box”, and users are also gradually being “datalized” under the in?uence of algorithm. Based on the above situation, this research, along the thinking of “technological moralization”, proposes several ways to solve the problems, including incorporating news transparency into the “routine” of algorithm design, strengthening algorithmic stakeholders’ democratic cooperation, and readjusting the relationship between human and algorithm through dialogue rather than confrontation.