We still don’t have a “please enter me into #gardeningtok” button, and TikTok still has questions about the content it recommends, but it’s nice to see TikTok giving users more control over their own feeds. For a while, the app has had a “Not Interested” button, which you can use to stop seeing videos from a specific user or those that use a specific sound. Adding hashtag filters is a logical next step, similar to what other social platforms have done.
Currently, the word-based filters only look at video descriptions and text-based stickers. But, in theory, they could become much more powerful in the future: captions are a core feature of TikTok, and its automatic captions are typically quite good, so the platform could begin to more deeply understand the content of its videos simply by having their transcripts. For the time being, you may only be able to filter out individual words, but as machine learning and language models improve, that feature may become more sophisticated. That same effort will benefit the rest of TikTok’s moderation efforts as well. The more it understands its content as well as its audience, the better it can tailor the experience. But neither is simple to understand.
