You open TikTok to see a quick video. But then 25 minutes have passed, and you’re still swiping. A cooking video gives way to one about travel, then an interview, a tech product review… it seems as though the social network knows precisely what will interest you.
That’s no coincidence, because TikTok’s algorithm is actually one of the most sophisticated recommendation systems on the internet, a technology designed to select and order videos that appear in the “For You” tab. Its influence is so immense that it can pick and choose which videos go viral, which subjects gain in relevance, and to a large extent, how millions of people consume information every day.
The secret: your interests
In contrast to other social media networks that take into account the pages we follow, showing us related content, TikTok is largely based on an “interest graph,” attempting to determine what we like by observing our behavior, not just social connections.
The company itself explained in 2020 that its “For You” page is created using several indicators. Among them are the videos we watch all the way through, what we replay, our “likes,” comments, content we share, the accounts we follow and even the videos we decide to hide or mark with “I’m not interested.” It also takes into account information about the videos that we upload, as well as the hashtags, descriptions and sounds we use. In other words, every second we spend looking at a video adds information to its system.
One of the unique aspects of its algorithm is that it assigns special attention to implicit signals, to actions we carry out without being completely aware. For example, watching a video all the way through to the end is often interpreted as a display of stronger interest than simply giving it a “like.” Playing it several times, or staying a few seconds longer than normal can also indicate that we find the content attractive. And, according to TikTok, these signs help to predict which videos may interest us in the future. That’s why many people have the feeling that the application “reads their mind.” In reality, the system is analyzing extremely detailed patterns of behavior.
How it molds what we consume
The recommendation algorithm doesn’t just show content: it also influences the way we discover information. When a person interacts repeatedly with videos about a specific topic — sports, politics, nutrition, technology, investing — the system tends to offer more related content. This can be useful, because it allows us to quickly discover publications that pertain to our interests. However, it can also reduce the diversity of the content we see.
TikTok itself recognizes that risk, and says that it deliberately introduces videos about subjects outside a user’s typical interests to keep the experience from becoming too homogeneous. It also says that it tries not to show two consecutive videos from the same creator or on the same subject.
Still, recent studies have observed that TikTok’s recommendation systems can reinforce specific interests quickly. In fact, a study published in 2025 by researchers from Cornell University found that the amplification of content aligned with a user’s preferences can take place during the first few hundred videos they consume and that through this personalization, our exposure to new subjects is reduced.
How to take back control
The efficacy of the algorithm has also raised concerns among regulators and experts. The European Commission is studying TikTok’s different design characteristics, including its infinite content scroll and high level of personalization of recommendations. Policymakers think such elements could lead to compulsive patterns of use, especially among minors and vulnerable users. TikTok rejects these accusations, and says it makes use of protective tools for digital well-being.
Although it turns out to be impossible to completely dismantle the algorithmic logic of the app, there are ways of reducing its influence. For example, being conscious that every interaction counts: staying longer on a video sends a signal to the system, as does sharing and commenting.
It’s also useful to search for new topics, follow different kinds of profiles and avoid constant interaction with the same kind of content. And it’s recommended that users utilize the management tools that the platform itself offers, like marking videos that don’t interest them or periodically reviewing its preference configurations.
Lastly, it’s worth a reminder that the algorithm is optimized to maximize relevancy and the time we spend on the platform, not necessarily offering us a balanced view of the world. Understanding that difference is probably the first step towards using TikTok more consciously.
Sign up for our weekly newsletter to get more English-language news coverage from EL PAÍS USA Edition
