This post was written by Shuaishuai Wang and Rachel Spronk as part of the GDC-funded project “From Performativity to Behavioral Data: The Algorithmic Configurations of Sexuality on Social Media in China“.
Meet Loki, a short video creator on Douyin, the Chinese version of TikTok. Documenting daily life with his same-sex partner with short videos, Loki has accumulated around 141,000 followers and more than 1 million likes on Douyin.
“Douyin’s recommendation algorithms are scandalously precise. This means that the platform bases its recommendations not only on general, cursory metrics such as the male gender, but on much more refined and nuanced subdivisions, including age range and body shape. These subdivisions exist in each category, and help pinning down your preferences precisely,” said Loki, in an interview with us in February 2021.
Loki’s story articulates a broader, shared experience among same-sex interested groups in China, which can best be captured by the phrase “big data sees through you” that went viral since 2019. The phrase refers to the phenomenon that Douyin is able to identify users’ sexual curiosities and orientations through gendered connections emerging from watching short videos. For example, if a viewer only engages with videos created by one particular gender, the platform’s algorithms will pick up on that and recommend same-sex videos to the viewer. On Douyin, “big data sees through you” has gained potency featuring men wearing basketball shorts, white socks/shoes, or men displaying their gym-trained bodies. Female users seldom show their bodies in short videos, they act “cool”, so to say, by dressing in a unisex way, or as what is considered manly, while only some women enact same-sex performances that are understood as intimate. Such particular, almost fetish, preferences are an important phenomenon in the Chinese digital queer cultures. Queer creators also play with the “big data sees through you” phrase in hope to reach their targeted audience that is (also) interested in same-sex intimacies (see the figure).
Users’ data on their online behaviors are central to Douyin’s recommendation algorithms. Those data emerge from collecting the metric of people’s interaction, such as the amount of time spent on a video, liking, commenting, sharing, and browsing the comment thread after viewing a video. These data are aggregated, correlated, and interpreted by the platform to establish a spectrum of interests from users. In doing so, Douyin reorganizes and simultaneously reinterprets users’ social networks and their content preferences and, further, on the basis of which it recommends videos as reassembled by Douyin.
In recognizing this “magic power” of Douyin’s recommender system, with which they mean the algorithms are so smart as to be able to identify users’ sexual interests, Douyin creators invented the phrase “big data sees through you.” It became a recognizable badge for same-sex interested people. Interestingly, one can hardly find terms such as gay, lesbian, or queer in these videos. Sometimes people use the shortened three letters “txl” to make their position as being interested in sexual diversity more recognizable. Txl stands for “tong xing lian”, the Chinese translation of homosexuality. In our interviews, seldom do these self-identified gay men or lesbian creators see problems of expressing their sexualities, or erotic interests, via big data analytics. On the contrary, they rave in the fun of it, without making references to the language of sexual identities.
In China, sexual identities articulated by and produced in the languages of LGBT+ (gay, lesbian, bisexual, transgender and more), as well as the global discourse on queer, have been highly politicized and therefore classified as sensitive. The algorithmically configured sexual and erotic classifications, on the other hand, are nascent and in constant flux because of the real-time data feedback loop. For these reasons, algorithmic identifications are often misread to be “technical” and thus “free of politics.” These qualities help create spaces for data-based sexually diverse cultures. In our project, which was funded by Global Digital Cultures at the University of Amsterdam, we aim to explore how algorithmic configurations of sexual classifications and identifications come about and how they may or may not provide opportunities for visibility and sexual politics at large. We are interested to investigate how algorithms make sexual curiosities, erotic preferences, and sexual identifications fluid and elastic, as opposed to fixed signifiers.
Shuaishuai Wang
Shuaishuai Wang is a Lecturer in New Media and Digital Culture at the Department of Media Studies. Focusing on TikTok/Douyin, his current research investigates the complex interplay of infrastructures, algorithms, creators/users, and platform companies in the global advancement of the digital culture industry.
Rachel Spronk
Rachel Spronk is Associate Professor at the Department of Anthropology. She works at the intersection of three scholarly fields – anthropology, gender & sexuality studies, and African Studies. In her work she combines the ethnographic study of practices and self-perceptions with the task of rethinking our theoretical repertoires.