Module 1-2: Game Recommendations

One fairly common example of AI today is recommendations. Recommendations could be for movies, TV shows on streaming services, books or video games on other websites. It's how YouTube knows that you like watching gameplay videos, how streaming sites like Amazon know you like dramas, and how music apps like Spotify know you really like hip hop. Recommender systems used to be very primitive, but can predict a shockingly large amount of things about you.

Let's look at the video game platform Steam, for example.

Much like any other site like Amazon, Steam will recommend games that it thinks the user will like. Many factors, from the user's shopping history to their playing habits, are taken into account. While a seemingly simple process, there is actually a lot of information about a given customer that is used to try and tailor the experience perfectly to them.

How does using AI actually generate recommendations, though? One of the primary ways is simply to take the data from a user's favorite game and recommend other popular games in the same genre. Someone who is a big fan of RPGs might like a popular RPG such as Final Fantasy 7. The more data you get, the more accurate recommenders can be.