Gone are the days when social platform algorithms were secret and almost magical formulas. TikTok is now rolling out a new feature that will let users understand why a certain video is being served to them in their For You feed. The new feature is designed to give you more context on how the For You feed works.
If you’re new to the platform, the For You feed is a similar section to Instagram’s Explore feed. This is the part of the app where the content that the algorithm believes may interest us the most, based on our interactions with other videos on the platform. The For You feed is the main section where TikTok users spend most of their time.
TikTok will reveal why a certain video appears in For You
The new feature will work very simply, and is actually already available to a select number of users in the US. Basically, just click on the share icon of a video that appears in the For You feed. Then u will appearn icon with a question mark labeled “Why this video”. By clicking on it, the app will explain why the algorithm thought that we might like that content.
Generally the algorithm of TikTok tends to recommend videos similar to content we’ve interacted with (through likes, comments and shares). Not only that: the AI also takes into account the time we spend on individual contents before swiping and move on to the next. In other cases, however, he could decide to advise us popular videos in the region where we are geolocated.
“This feature is just one of many tools we are working on to provide greater transparency for those who use our platform,” the company said in a blog post. “Looking forward, we will continue to expand this feature to provide more granularity and transparency around recommended content.”
The For You feed has recently been at the center of a controversy following a study conducted by the Center for Countering Digital Hate (CCDH). The researchers found that the feed displayed dangerous content every 206 seconds on average. The videos in particular concern eating disorders and self-harm. We talked about it here: