An Empirical Investigation about awareness of personalized recommendation algorithm among Tiktok users: An Actor-Network Theory Perspective


Personalized recommendation system has been developed for more than 20 years and has been widely used. People are becoming increasingly reliant on online socio-technical systems that employ algorithmic curation to organize, select and present information. Personalized videos and music can enhance people's preferences, which may also lead people to indulge in pleasure. People's ignorance of algorithms may be threatened by platform and algorithm control, so we need to study users' awareness of personalized recommendation algorithm. Up to now, there have been few studies that examined users’ awareness of personalized recommendation algorithm in China. Based on the theory of actor network, this paper takes domestic Tiktok users as the research object to explore users' awareness of personalized recommendation algorithm. Because short video is the second-largest Internet application in China after instant messaging, it has more users than search and online news. Watching short videos has become an important part of Chinese netizens' daily life. In particular, Tiktok app not only has the largest number of users and rich UGC, but also has the leading personalized recommendation algorithm technology, and users have a lot of discussions on personalized recommendation algorithm.

To achieve the goal, I focus on two sites of investigation. In the first part of the study, I examine the extent to which Tiktok users have been aware of personalized recommendation algorithm. It shows that many Tiktok users(N=203) recognize personalized recommendation algorithm. According to the collected data, 75% users have heard of personalized recommendation algorithm, 49% users said they knew about personalized recommendation algorithm while their knowledge is limited. Moreover, 63% users have a low desire to know personalized recommendation algorithm, many people think it doesn't count so much.

The second part explores users' specific understanding of personalized recommendation algorithm. First, I introduced the principle of six personalized recommendation algorithms, and also analyzed the specific algorithm design of Tiktok platform. Tiktok platform has a strong paternalistic style, and users feel the strict algorithm rules. Then, through a thematic analysis of algorithm discussions among Tiktok content creators, it is observed that although they are outside the context of traditional technical expertise, they have developed some algorithmic knowledge through practice and conducted the act of “gaming the system”. Instead of fully embracing the platform's personalized recommendation algorithm rules, they have cultivated some expediencies to circumvent the platform’s set rules, with lots of metaphorical language observed(e.g., “raise the smurf account”、“ dig the grave”).Moreover, I summarize the anthropomorphic concept of the algorithm. In their eyes, the algorithm acts as content auditor, attention distributor, judge and referee. Thirdly, I employ the Actor-Network Theory to research and analyze the various actors around the topic of personalized recommendation of short videos. On the whole, users' understanding of personalized recommendation algorithm is practical, and most people show indifferent attitude. Although there are acts of “gaming the system”, they are still catering to the algorithm, without strong resistance and criticism. Finally, this paper makes a critical reflection on the ideology of network individualism and inequality about personalized recommendation algorithm.