Anti-You

For-You Page
Anti-You Page
Settings
Filter Video Topics

Inspiration

People can spend hours scrolling on TikTok through the For-You page to be immersed in their own subcultures, and it can be very hard to break out of their own bubble. Not only is it hard to escape this algorithm, but it limits the opinions and content you interact with. Personally, we've had to make new TikTok accounts either because we outgrew the content or felt isolated within what TikTok had determined were our favorites. Having a page cultivated for you specifically can perpetuate confirmation bias and not let you question your own beliefs. This is very important in this age of misinformation on the internet. We want to break the norm for each person by developing an anti-For-You page.

What it does

What is an Anti-You page? An Anti-You page is a separate page that a user scrolls through like any other page on TikTok (For-You, Following, STEM, etc.). However, the page will recommend TikToks the user would not normally expect. We use historical user data to determine the content they have least interacted with, along with AI to infer more content they are not likely to interact with and suggest popular videos under these categories in their Anti-You page.

Prominently, interacting with TikToks on your Anti-You page can influence the content shown on your For-You page. Essentially, its an easier way to break your own bounds. However, we have offered to make this a Setting, such that a user can decide if they wanted their For You page to be modified by interacting with Anti-You.

Finally, we've offered filtering features in the Settings, such that users can filter their Anti-You pages based on categories they want to explore. This includes adding or removing categories to ones already determined by our algorithm .

How we built it

We used Figma to create a prototype of how we envision this feature to look like. We expect this to be an extension to TikTok, and the company idea would generally be to sell our feature/algorithm to TikTok, and in the future, other algorithm-based methods of content curation (such as Instagram's explore page)

Challenges we ran into

Deciding how much agency to give our users: giving users too much agency in terms of filtering categories can defeat the whole purpose of suggesting content to them that they are least likely to interact with.
Figma: designing our prototype such that we don't clutter the app.
Defining "least interacted" + using data analysis to clean and query the data for this information.