WSJ report on user ‘clusters’ reveals how unsettling TikTok’s data collection can be

It’s already known just how extensive TikTok is about collecting user data. From names, users’ approximate locations, and IP addresses, to even keystrokes(opens in a new tab), it’s pretty intense. This meticulous level of data collection is an advertisers’ wet dream, but it’s the method by which TikTok catalogs that data that reportedly gave rise to recent internal concerns.

As The Wall Street Journal reported(opens in a new tab) on Friday, former TikTok employees have claimed that the app has been tracking the videos that users watch under topics, including “LGBT,” and in essence compiling lists of users who watch such content, which could at one point be viewed by some employees through a dashboard.

The WSJ described the groupings as “clusters,” which function like the notorious “taste clusters(opens in a new tab)” at Netflix that have been extensively mocked and parodied(opens in a new tab). They had names such as “mainstream female,” “alt female,” “southeastern black male,” and “coastal, white-collar male,” according to the Journal‘s report. TikTok does not ask for users’ sexual orientation, but based on the content users watch, it appears the algorithm was, at the very least, assuming that users were members of the LGBTQ community and categorizing them accordingly, all in the name of getting people to use the app more.

In one telling example, the WSJ notes that the “alt-female” cluster branches out into content related to “tattoos, some lesbian content, and ‘Portland.'”

As noted in the report, it is not uncommon for many social media and ad-tech companies to infer traits about their users based on online behavior. They use it to select which content or ads to serve to users. However, with TikTok’s “clusters” system, liking LGBT content didn’t just mean you were shown more queer-friendly content, but it seems the app as a whole essentially labeled users as members of the community.

This sketchy way of cataloging user data lead to internal worries, according to the WSJ, as some TikTok employees could view the unique identification numbers of users and the list of users who were watching videos in each cluster. This raised fears among the workers that the data could be shared with outside parties or used to blackmail LGBT users, the WSJ reported. Especially since Tiktok has admitted in the past to spying on reporters.

A spokeswoman for TikTok told the WSJ that the app doesn’t identify sensitive information based on what users watch and that users’ interests do not necessarily represent their identity. TikTok also confirmed that the dashboard used to access data on watchers of gay content was deleted nearly a year ago.

TikTok still collects this data but has simply replaced the cluster names with numbers and restricted access to a smaller number of employees within the company’s new U.S. unit.

It’ll be interesting to see how this new development plays out as the US continues to push to ban TikTok in the country. However, given the US government’s own capabilities of spying on its own citizens and its, at best, “tolerance” of the LGBTQ community right now, this might not mean much to the relevant policymakers in DC.