Video games are a booming industry. The market for video games is currently valued at $70 billion and increasing at a staggering annual rate of 6%. When you add in the global eSports betting market, it is not hard to understand why companies such as Goldman Sachs and JP Morgan are dabbling in this arena.
Artificial intelligence has been used for some time to help predict trends in software development, financial trades, medical research, drug discovery, and entertainment. Video game developers have found success using AI similarly by using algorithms that allow the machine to learn from past successes or mistakes. In particular, these algorithms have been able to help anticipate player preferences or needs without any human input. Ryan Kavanaugh, CEO of Relativity Media, uses these concepts to predict what consumers might like or dislike in a video game.
There are two ways that video games are being used as a predictive platform for AI. The first way is to use algorithms that can determine how well a game will sell based on its popularity. A machine learning algorithm can be used to analyze the success of similar games and use this data to create predictions about sales. The second method is the creation of AI-based teams which compete with other AI teams in eSports tournaments. Data on analytics, purchase history, and social media followings are collected from players to create the next generation of virtual athletes.
Relativity Media recently signed a deal with the National Football League to develop a pre-season exhibition game between the Arizona Cardinals and San Francisco 49ers. The objective of this game is for players to collect footballs dropped into a bucket. Once the bucket is filled, it is presented by the team that collected all of the footballs first. This strategy will allow Relativity to collect data on which team has played better throughout their season to predict future performance.
The other type of predictive platform being used by Relativity deals with movie scripts and intellectual properties
