The Filter Bubble

Eli Pariser used to follow people with a range of opinions on social media. He was a progressive activist, but he also liked reading what his conservative acquaintances thought. One day, he noticed that their posts had disappeared from his feed, without warning or explanation. It might seem like a small detail, but this kind of silent adjustment happens millions of times a day around the world.

What happened was simple. The social network's system noticed that he clicked more often on links from people who thought like him and started showing him only that kind of content. No one decided to erase the others. The algorithm simply cut out what seemed less interesting.

That observation led to this book, published in 2011. Its argument is that the personalized internet puts us inside a custom-made bubble, which he called a filter bubble, and that this bubble changes what we know about the world without our noticing.

The central issue is an invisible filter: the content you see has already gone through a selection process you never requested and know nothing about.

The internet began with the promise of opening doors. Anyone could search for any subject, access diverse sources, and talk with people around the world. The promise was that barriers to information would come down.

But in the years that followed, companies discovered that the money was in knowing who each user was. The more they knew about your tastes, the more precisely targeted ads they could sell. Personalization became the dominant business model. It worked so well that practically every digital service adopted it, from ride-hailing apps to music platforms.

Today, this seems natural. The online store suggests products similar to ones you viewed, the video platform recommends the next episode, and the search engine adjusts the order of results based on your history.

Pariser doesn't condemn convenience. He wants to show its other side: a world organized around what you already like tends to become increasingly narrow.

One example in the book involves two people who type the same word into a search engine and receive different results. In one case he describes, a search for a major oil company brought one person reports about an oil spill and the other information about investing in the company.

According to the author, the search engine uses dozens of signals to decide what to show, such as your location, the device you use, and your previous searches. In 2009, the company announced that it would begin personalizing results for everyone, even people who weren't signed into an account.

The result is that there is no longer one first page that looks the same for everyone. What appears depends on who you are, or who the system thinks you are. This difference is subtle and almost impossible to notice on your own, because no one sees their neighbor's screen.

This helps explain why two people in the same family can find very different news stories when searching for the same topic and come to opposite conclusions.