Recommender Systems Do Nothing For You

There once was a time where going viral on the internet actually meant something.

Long ago, in the early 2010s, 500,000 views could actually land you on daytime TV, where you could experience the brevity of fame before returning to a life of routine American living.

Internet culture previously existed as a monolith. Trending videos made their rounds across multiple platforms and there was a general understanding that once something was officially “viral,” most people saw it. 

However, a shift in recent years has changed the idea of social media being a series of relationship focused platforms into more of an attention driven entertainment ecosystem.  

Most platforms believed to be social media apps are actually entertainment apps, completely driven by recommender systems. 

Recommender systems are known by different names like the “the algorithm” or a “for you page.” Even Netflix’s “because you watched this movie” section is a form of a recommender system. User-generated content is the new frontier. 

The results of recommender systems are digital space bubbles. Recommender systems keep track of a user’s input — what is watched the longest, what is liked most often and what is engaged with most. Not only do they show more of the same content, but they only expose users to others who also like the same content. 

Entertainment is now highly personalized. Instead of one monolithic digital culture, we have personalized echo chambers, bouncing our own interests and beliefs back at us to keep us scrolling. 

While this is profitable for advertisers and marketers in selecting an accurate target audience, it has negative side effects. 

Highly niche communities have always existed on the internet, but with the age of recommender systems, they’re not only becoming more prevalent, but far more radical.

An example of this is the rise of extremist groups on social media. With the help of recommender systems, a user can fall down the rabbit hole into an extremist viewpoint just by not being exposed to any other opposing ideas or dissenters.

When a user is only shown content which justifies their point of view, it only solidifies the feeling their opinion is correct. In less than a decade, recommender systems have helped divide and polarize users at a speed never seen before. 

Internet culture is no longer a monolith, but an us-versus-them whirlpool that can suck even the youngest, newest internet users in. 

The ironic aspect of this horrible situation is the internet’s original purpose was supposed to democratize information — giving anyone around the globe access to a variety of articles, opinions and other media. As long as someone had a device and internet access, they could be given exposure to any perspective, any news outlet and any social media platform. 

Recommender systems shut us into our own bubbles, actively keeping us from other worldviews. 

It can start small — through one like, one follow or one repost. It’s enough for your recommender system to latch onto and begin closing a user into a specific niche, despite there being more content available than ever before in history.

There is a way to maneuver out of an echo chamber through not consuming media passively and instead intentionally exploring opposing viewpoints. Recommender systems love to validate pre-existing ideas to keep users entertained and scrolling. Even if one doesn’t agree with that viewpoint, it will widen a closed perspective and give an opportunity to remove one from a bubble.