A new study conducted by computer scientists at Northwestern University found that TikTok's algorithm for content recommendations responds to user feedback, such as the "not interested" button, but only for a temporary period. The algorithm gradually reverts to showing previously excluded content unless feedback is persistently given by the user.

The study indicated that using the "not interested" button reduced unwanted content by approximately 84 percent. Merely skipping videos resulted in a 48 percent reduction in unwanted content. Levi Kaplan, a Northwestern University computer scientist, said, "So if you don't want to see something, you should be hitting that button." Piotr Sapiezynski, also a computer scientist at Northwestern University, said, "When you start saying, 'I don't want to see this particular topic,' the platform might actually show you fewer of such pieces of content."

However, the study also determined that even brief re-engagement by a user with previously unwanted content is enough for the algorithm to relapse. "It turns out that it works in the beginning," Sapiezynski said. "But then the platform will slowly start putting it back in your feed." He added that if users do not continue to express disinterest, the content may return to its original frequency in their feed. "And if you don't continue saying, 'I really don't want to see it,' this may balloon back to the place where it was in the beginning," Sapiezynski said. He stated that the platform reacts to negative feedback initially, but also to user behavior. "So the platform does react to your negative feedback, but then it also very much reacts to your express behavior," he said. "So if you are presented with this content again and you start watching it, the platform will again feed it to you more and more."

The researchers used bot accounts on the actual TikTok mobile app for their methodology. Kaplan explained, "We used emulated devices, where we are creating accounts and automatically interfering with the TikTok algorithm through code with the sock puppet accounts." The team ran experiments multiple times on 90 cloned accounts, focusing on three topics: cooking videos, fitness videos, and sports betting. The researchers obtained metadata by intercepting network traffic and made decisions using large language models (LLMs), which were validated with human responses.

The authors of the study suggest that the "not interested" option appears to be deliberately concealed from users. Sapiezynski said, "On the other hand, it's unclear why the platforms would offer it, if it doesn't work." Kaplan stated, "We can teach users how to use the platform better, but ultimately the way that you're interfacing with the platform is going to be dictated by the design decisions that are fundamental to the platform." The study was published in the Proceedings of the Twentieth International AAAI Conference on Web and Social Media, 2026. The researchers plan to test their hypothesis on real user data in the future.

Why It Matters

TikTok's For You Page (FYP) serves as the default home screen for users, and its algorithm utilizes both implicit signals, such as video watch time, and explicit signals, like likes or follows, to determine content recommendations. The study's findings indicate that while users have some immediate control over the content they see, this control requires consistent and repeated effort to maintain. The ineffectiveness of a single interaction to permanently alter content suggestions could impact user experience by requiring continuous engagement with feedback mechanisms.