Youtube’s Recommendation Algorithm
https://www.scientificamerican.com/article/youtubes-recommendation-algorithm-has-a-dark-side/
Youtube makes its money through targeted ads on their videos. Youtube wants its users to stay on their site longer so they can watch more videos, making them more money. Youtube influences its users to stay on their site longer by recommending them videos that may pique their interest. Using a specialized algorithm, Youtube is able to find and recommend videos to users that they might click on. Clicking on one of these videos leads users into a wormhole of similarly related videos that appears almost endless.
One of the ways Youtube might find these videos is through the usage of strongly connected components. When users are interested in one type of video, they may also be interested in other videos relating to that subject. Strongly connected components can be formed showing related videos that users with similar interests might enjoy. Youtube’s recommendation system takes advantage of this, so once a user watches one or two videos from a new category, Youtube will recommend the user other videos that are related to the video they watched.
Although this algorithm can be useful towards keeping users engaged on the site, it can lead to some moral problems. Using strongly connected components to recommend videos can lead to the creation of echo chambers. If a user becomes interested in a conspiracy video about Flat Earth theory, for example, they will be shown more videos about Flat Earth theory, only confirming their incorrect belief. They will not be shown any videos denying the theory because they are less likely to click on it.