Is TikTok a time bomb?

In its meteoric rise and its approach to leveraging user data, TikTok may augur the AI-informed future of social media, for better and for worse.

You may have never heard of TikTok, but chances are the kids in your life have. The China-based social network made its U.S. debut in 2017 and now counts 104 million American downloads (1.2 billion worldwide). Since January, it has consistently ranked among the top three most downloaded apps, behind Facebook-owned duo WhatsApp and Messenger. The average user now spends around 45 minutes a day on TikTok, more time than they even spend on Facebook.

What’s the appeal? On its surface, TikTok is pure social media candy that’s highly addictive to its tween audience, according to user data analyzed by App Ape. Users make 15-second videos (think funny dances, pranks, and lip-syncing bits), then use TikTok tools to incorporate music and effects. (Here’s what’s trending today.) They can then publish their creations to the whole world, YouTube-style. One of the best-known TikTok stars is Lil Nas X, whose Billboard-topping hit “Old Country Road” got its start on the platform.Beneath the surface, however, the picture is more complex. In its meteoric rise and its approach to leveraging user data, TikTok may augur the AI-informed future of social media, for better and for worse.

BEHIND THE MEMES

TikTok’s parent company is hardly as warm and fuzzy as the platform would suggest. Valued at $75 billion, China-based ByteDance owns a suite of apps built around AI and fueled on user data. Among the first Chinese platforms to make significant inroads in the U.S., ByteDance is fast becoming a social media giant. It has a staff of 40,000 (roughly 10 times as many employees as Twitter) and has expressed serious interest in buying Twitter and Snapchat.

TikTok’s rise hasn’t been without controversy. In February, it was fined $5.7 million by the FTC for collecting personal information from kids under 13, violations that occurred inside a platform it acquired in 2017 called Musical.ly. (The day after it paid that fine, TikTok added online safety videos and comment filters to its app.)

Meanwhile, TikTok’s youthful audience has drawn unwanted attention from sexual predators. Authorities in India have gone so far as to ban the platform outright. At a moment when U.S.-China trade tensions are simmering, questions have also circulated around data surveillance and censorship from the Chinese government.

But what sets TikTok apart isn’t necessarily concerns about privacy or predators. All platforms wrestle with these issues. Rather, it’s the way it has risen to prominence—on the back of sophisticated AI and nearly limitless ad money. ByteDance spent roughly $1 billion in ads (as much as $3 million a day) in the last year, much of that for ads that appeared on the traditional social networks, according to the Wall Street Journal.

In contrast to earlier channels like Facebook and Twitter that grew organically (if exponentially), TikTok has bought its way to the front of the queue. Even its trending memes are largely manufactured. Before its U.S launch, the company hired a virtual army of social media celebs to create content for the platform—paying one influencer more than $1 million for a single video.

While ads have brought TikTok exposure, its groundbreaking AI is what keeps users hanging around. According to Bloomberg, “Within a day, the app can get to know you so well it feels like it’s reading your mind.” ByteDance founder Zhang Yiming is regarded as something of an AI savant inside China, with an uncanny ability to build platforms that harness user data to addictive effect. His first major hit, a news app called Toutiao, built profiles comprised of as many as 2,000 discrete tags used to drive personalized content, growing to 700 million users. Douyin, the app that would become TikTok outside China, was launched in 2016 and made use of similar technology.

Significantly, TikTok’s recommendation engines are distinct from algorithms used by Facebook and traditional networks, which rely heavily on suggestions from friends. Rather, TikTok is like Toutiao in that it harvests insights based on what its users actually click on, read, and watch—right down to the type of music, faces, and voices in videos—learning as it goes. The distinction is subtle but important. Facebook was revolutionary because it tapped into our social grid of friends to serve relevant recommendations, but TikTok goes right to the source using AI to map out interests and desires we may not even be able to articulate to ourselves.

AN AI FUTURE FOR SOCIAL MEDIA?

What does this mean for the future of social media? On the one hand, TikTok’s algorithms showcase the true potential of AI and machine learning and may well have the traditional social networks worried. (The Wall Street Journal reports that Snap debated whether to sell ads to TikTok at all, for fear of losing users to a new rival.) ByteDance has already hired away executives at Facebook and is now turning its efforts to attracting older users. Its growth shows little sign of letting up.

But TikTok’s ascent comes at a moment when user privacy and the weaponization of data has been thrust into the spotlight. As the recent outcry over photo filter FaceApp illustrates, users are beginning to appreciate just how much valuable personal data they give away often for little more than a passing moment of entertainment. TikTok may well epitomize this “cheap date” syndrome. Target revenue for 2018 was estimated to be around $7 billion, largely from selling highly targeted ads based around mountains of free user data. Whether increasingly savvy users will continue to abide with this commercialization of their privacy, only time will tell.

State surveillance presents an even larger concern. It’s a truism that private corporations in China and the Communist Party are intimately linked. TikTok’s domestic equivalent, Douyin, is thought to be closely monitored by authorities. The same advanced algorithms that monitor and pick up on user preferences also make it easy to identify, for example, a traditional Uighur song sung in the native language and to alert authorities. There’s no evidence that surveillance capabilities have been applied outside China, but the possibility has been raised.

Whether all of this will actually impact the growth of TikTok and attitudes toward the platform remains to be seen. Despite the simmering backlash, viable alternatives to existing monetization models—whether in the form of ad-free subscription social media services or data marketplaces where users opt to share their data for a price—have yet to get off the ground. But what is clear is that with the rise of TikTok, the clock is ticking. AI and user data are converging in complex and sometimes unpredictable ways. Will we as users be able to keep pace?

Ryan Holmes (@invoker) is the CEO of Hootsuite, the social media management platform with 18-million-plus users.