Many people and researchers have tried guessing and finding out how the TikTok algorithm is so good, but some of the results that have been revealed in the past have left some people and media disturbed and amazed at the time.
A former Google engineer who worked on YouTube’s algorithm has revealed that the platform uses a very similar, essentially the same method, although it’s not as extreme as TikTok’s (via 9to5Mac). TikTok uses four signals to determine what you watch, share, like, and follow. A recent analysis revealed that having one of these is already more than enough to get to know a user’s favorite topic and main interests, but having the three others provide even more help narrowing down and making the algorithm better.
The Wall Street Journal created a 13-minute long video to share its findings, but we’ve collected some of the important and interesting bits here.
“We found out that TikTok only needs one of these to figure you out: How long you linger over a piece of content.
Every second you hesitate or rewatch, the app is tracking you. Through this one powerful signal, TikTok learns your most hidden interests and emotions, and drives you deep into rabbit holes of content that are hard to escape.
The TikTok experience starts the same way for everyone. Open the app and you’ll immediately see an endless string of videos in your For You feed. TikTok starts by serving the account a selection of very popular videos vetted by app moderators.”
The WSJ has created 100 bot accounts with an age, location, and a few interest points to conduct these tests. What’s interesting is that the interests were never entered directly into the app, it simply picked videos and watched them based on the hashtags and content identified. The bot was programmed to search for hashtags that the AI would associate with the interest set for that particular bot. When the bot found a video that matched its interests, it stopped scrolling and watched the content. Sometimes, it rewatched the video multiple times to ensure the algorithm knows the bot enjoyed it.
The research found that all of the videos and view counts were contracted and tightly focused on the user’s interests.
“The results were analysed by data scientist and algorithm expert Guillaume Chaslot, a former Google engineer who worked on YouTube’s algorithm.
He’s now an advocate for algorithm transparency. He says TikTok is different from other social media platforms.
“The algorithm on TikTok can get much more powerful and it can be able to learn your vulnerabilities much faster.”
In fact, TikTok fully learned many of our accounts’ interests in less than two hours. Some it figured out in less than 40 minutes.”
WSJ programmed another bot to have sadness and depression interests.
“Less than three minutes into using TikTok, at its 15th video, [bot] kentucky_96 pauses on this [sad video about losing people from your life]. Kentucky_96 watches the 35-second video twice. Here TikTok gets its first inkling that perhaps the new user is feeling down lately.
The information contained in this single video provided the app with important clues. The author of the video, the audio track, the video description, the hashtags. After kentucky_96’s first sad video, TikTok serves another one 23 videos later – or after about four more minutes of watching.
This one is a breakup video with the hashtag #sad. TikTok’s still trying to suss out this new user, with more high view count videos [but] at video 57, kentucky_96 keeps watching a video about heartbreak and hurt feelings. And then at video 60, watches one about emotional pain.
Based on the videos we watched so far, TikTok thinks that maybe this user wants to see more about love, breakups, and dating. So at about 80 videos and 15 minutes in, the app starts serving more about relationships. But kentucky_96 isn’t interested. The user instead pauses on one about mental health, then quickly swipes past videos about missing an ex, advice about moving on, and how to hold a lover’s interest. But kentucky_96 lingers over this video containing the hashtag #depression, and these videos are about suffering from anxiety.
After 224 videos into the bot’s overall journey or about 36 minutes of total watch time, TikTok’s understanding of kentucky_96 takes shape. Videos about depression and mental health struggles outnumber those about relationships and breakups. From here on, kentucky_96’s feed is a deluge of depressive content. 93% of videos shown to the account are about sadness or depression.“
The following is proof that the algorithm can easily be manipulated by watching a topic or interest for a certain amount of time. Once the algorithm finds these interests, it’ll automatically adjust and include only videos in the feed that matches the user’s interests.
This method works great for funny and uplifting content, but much less so for those who are suffering from depression and sadness, as it can make them feel worse. 9to5Mac points out that conspiracy theorists could end up with the impression that such views are mainstream and widely accepted, even though some of the things are not fact-checked and proven to be true at all.
TikTok prioritizes engagement over mental health, and Chaslot says that YouTube does something similar with its recommendation algorithm, but it’s much less extreme.
What are your thoughts about the new findings of how TikTok works? Did the algorithm figure out your favorite interests and hobbies? Let us know in the comments!
Roland is a technology enthusiast and software engineer based in United Kingdom. He is also a content creator and writer, and is best known under the name “Techusiast”.