The company built a new subsidiary to house U.S. data, hoping to eliminate the concern that TikTok U.S. users / WSJ. For at least a year, so...
 |
| The company built a new subsidiary to house U.S. data, hoping to eliminate the concern that TikTok U.S. users / WSJ. |
TikTok doesn’t ask users to disclose their sexual orientation, but it cataloged videos users watched under topics such as LGBT, short for lesbian, gay, bisexual and transgender, the former employees said. The collection of information, which could be viewed by some employees through a dashboard, included a set of affiliated users who watched those videos, and their ID numbers, they said.
Other topics in TikTok’s data set also included lists of users, but the former employees didn’t consider those topics to be sensitive. TikTok workers in the U.S., U.K. and Australia in 2020 and 2021 raised concerns about this practice to higher-level executives, saying they feared employees might share the data with outside parties, or that it could be used to blackmail users, according to some of the former TikTok employees.
Many social-media and ad-tech companies infer traits about their users based on online behavior. They use it to select which content or ads to serve to users. Social-media and ad-tech industry practices, however, discourage tracking potentially sensitive traits such as sexuality, according to people who work with digital information. This data can essentially create a list of vulnerable users in parts of the world where some LGBT people face harassment and violence.
TikTok is in an unusual situation when it comes to its handling of user data. U.S. lawmakers critical of TikTok say they fear the Chinese government could force the app’s owner, Beijing-based ByteDance, to divulge information about its U.S. users. A spokeswoman for TikTok said TikTok hasn’t been asked, nor has it provided any U.S. user data to the Chinese government.
The spokeswoman for TikTok said the dashboard that employees used to access the data on watchers of gay content was deleted in the U.S. nearly a year ago. The spokeswoman said TikTok doesn’t identify potentially sensitive information such as sexual orientation or race of users based on what they watch. She also said TikTok doesn’t infer such information.
The data represent users’ interests and isn’t necessarily a sign of someone’s identity, the spokeswoman said. Users who engage with LGBT content on TikTok may not identify as LGBT themselves, just as there are people who enjoy baking content but aren’t bakers, the spokeswoman said. “Safeguarding the privacy and security of people who use TikTok is one of our top priorities,” TikTok said in a statement.
Internally, some employees had argued that the data was safe to collect because it didn’t indicate whether users really were members of a particular group, some of the former employees of TikTok said. Other employees disagreed, saying that topics of videos a user watched was often enough to infer aspects of their identity, particularly for topics such as sexuality, the former employees said. Those employees described the data as a list of TikTok’s users who are gay.
The Biden administration has demanded that TikTok’s Chinese owners sell their stakes in the video-sharing app or face a possible U.S. ban. Many federal agencies and local authorities have barred use of TikTok by government employees, and Montana lawmakers have approved a statewide ban on use of the app by anyone in the state.
TikTok has tried to be more transparent about how it handles data-gathering and content moderation, and TikTok built a new subsidiary to house its U.S. data that it says should eliminate the concern that TikTok U.S. user data could be subject to Chinese law. The company has said that a forced sale won’t address perceived security risks.
Former TikTok employees who have worked elsewhere in the tech industry say the dashboard that provided access to view the data was accessible to more workers than is common at other tech platforms. Employees in China also had access to the data, and at times controlled the permissions for who could view the information, according to the former employees.
The TikTok spokeswoman said that at the time, the company had protocols to ensure data was accessed only by authorized employees, including those in China. Now, she said, TikTok stores the data in its new U.S. subsidiary, and only approved employees in the unit have access to it.
Social-media platforms including Meta Platforms also track similar data, but in recent years have locked down access to sensitive information. No comprehensive U.S. privacy law regulates the practice of collecting sensitive data. Seven states have passed privacy laws, including some that require companies to protect certain categories of sensitive data, including gender and sexual identity.
The Network Advertising Initiative, an industry trade group that represents many players in the digital advertising technology industry, has since 2015 forbidden members from targeting people based on inferred LGBT identity. The trade group feared that targeted advertising had the power to inadvertently out a young person to their parents or peers on shared computers.
Leigh Freund, chief executive officer of the group, said members “have been willing to forego revenue opportunities for tailored ads to LGBTQ audiences” to avoid potentially outing people. Klon Kitchen, a senior fellow at the conservative American Enterprise Institute and a former intelligence officer said TikTok’s current ownership structure, and China’s national security law that requires companies to comply with government requests for data, pose unique risks to U.S. privacy and national security.
“If an individual isn’t public about that orientation then it theoretically could be used to embarrass them,” Mr. Kitchen said. Or, for the Chinese government, “they can use that in the context of manipulation,” he said. China’s Foreign Ministry has criticized the U.S. for targeting TikTok and said Beijing would never require companies to illegally gather data and intelligence from overseas.
At TikTok, the company organizes all the videos its users post into a web of clusters, sorted by topics, the former TikTok employees said. The clusters span the universe of TikTok videos, including ones named: mainstream female, alt female, southeastern black male, and coastal, white-collar male. Each cluster includes subgroups; for alt female, those included tattoos, some lesbian content, and “Portland.” A cluster about professional basketball, for example, had subgroups about the Golden State Warriors, and star player Steph Curry.
TikTok tracked the categories of content and users on its app in an effort to understand trends and find ways to boost engagement, some of the former employees said. Some TikTok employees could view the unique identification numbers of the users associated with each cluster, as well as the list of users who were watching videos in each cluster.
Additionally, employees could look up users based on their ID number—a series of numbers each TikTok user is given when they start watching videos on the app—to see what cluster they were associated with, according to some of the former employees.
TikTok executives became concerned that this data was too widely available within the company, and in 2021, TikTok restricted access to the dashboard, reducing the number of employees who had access to it, according to some of the former employees. Some workers had also grown uneasy with how TikTok employees were describing the clusters, and debated changing the names, according to some of the former employees.
TikTok still collects data about the types of videos on the app and the users who watch them, but in 2021 it opted to remove the names of the clusters, and replace them with numbers. In 2022, TikTok deleted the dashboard that employees had used to access the data, and moved the data to the company’s new U.S. unit, where a smaller number of authorized employees can access it.