从热搜榜看平台算法传播公共性建构的三重困境

The Triple Dilemmas of Platform Algorithm Dissemination of Publicness Constructed Through the Trending Topics

当下,算法正以不可见、不可知的方式运作,并持续影响着公共空间和舆论生态。本文通过对平台热搜榜的细致考察,发现由平台算法传播所建构的公共性存在三重困境:其一,多重行动者博弈下的算法价值设定是以经济利益为导向的,平台热搜榜的议题呈现出浓厚的社交化和泛娱化属性,行动者们关于公共性的“问题”和“兴趣”指向并不一致,热搜榜建构的动机与最终结果均呈现非公共价值指向;其二,平台集体是一种由数据痕迹生成的用户集合,这种集合是趣味用户的归类,并不能代表现实中的大众,依据数据痕迹经由算法计算化、定制化的平台热搜榜,无法代表公众观点与意见,其中的议题建构是被区隔和伪公共性的;其三,在公共议题建构中,个性化推荐算法和集体化热点算法是一对相互矛盾的算法信息导控机制,它们在平台热搜榜中分别形成“兴趣循环递归”和“议题循环递归”,进而导致公共性供给窄化。最终,平台算法传播造成选择性偏见、情感与审美茧房以及数字化、结构性不平等等反公共性后果。

Nowadays,algorithms are operating in an invisible and unknowable manner,which continuously affecting public spaces and opinon ecology.Through a detailed examination of platform trending topics,this article finds that triple dilemmas in the publicness constructed by platform algorithm:firstly,the algorithm value setting is guided by economic interests,and the platform trending topics exhibit strong social and pan-entertainment attributes.The actors'problems and interests regarding publicness are not consistent.The motivation and final results of the construction of the trending topics are both present non-public value orientation.Secondly,platform collective is a collection of users generated by data traces,which is a classification of interesting users,but cannot represent the public in reality.Based on data traces,algorithmically calculated and customized,platform trending topics cannot represent public opinions.The topic is distinction and pseudo public.Thirdly,personalized recommendation algorithm and collectivized trending algorithm are a pair of contradictory information guidance mechanisms.They form interest and issue cycle recursion in the platform's trending topics,leading a narrowing of public supply.Ultimately,the dissemination of platform algorithms results in anti-public consequences such as selective bias,emotional and aesthetic cocoons,as well as digitization and structural inequality.