chapter one

1 Recommender systems in action

 

This chapter covers

  • Recommender systems (RecSys) and their primary types
  • An introduction to algorithmic amplification
  • The main kinds of social media and how amplification manifests itself

Consider this unfortunate but common scenario: Your cousin, who has always been a little quiet and shy, starts sharing more and more posts on social media. At first, it’s just a collection of images with poorly drawn cartoons (or memes), including a smirking green frog that you dismiss without thinking much about it. After a while, however, you notice that the content becomes increasingly troubling, rapidly shifting to the extreme of the current political discourse, and even actively pushing it. The content starts becoming downright disturbing because your cousin seems to believe that a secret cabal of baby-eating individuals is running the government. His mistrust of public institutions grows, and he stops accepting public health recommendations, convinced they serve to keep the cabal in charge. He is no longer welcome at family functions, and he is becoming more isolated, to the point that you worry he might do something dangerous.

1.1 Recommender systems are everywhere

1.1.1 A world without RecSys

1.1.2 What exactly are RecSys?

1.2 Social media: From bulletin boards to doom scrolling

1.2.1 The explosive success of social media

1.2.2 RecSys in social media

1.3 Societal impact

1.3.1 The influence of platforms

1.3.2 Recommendation as amplification

1.3.3 Measuring recommender systems

1.3.4 Algorithmic amplification as a measure of societal impact

1.4 Summary