Video games are being used so AI can learn faster
Even though artificial intelligence (AI) and machine learning technologies are making such advancements that are now able to beat humans in competition, they still need to learn how to do everything humans do. It is true that the core of AI is the data processing and loading of all the possible scenarios of human action in a specific activity, and that is what scientists are aiming for right now. To use available resources to teach AI devices different skills like through the use of video games that can simulate the desired skill to be learned, such as driving cars or playing poker.
The video game industry has been functioning since the beginning with AI that has been developing to more complex fields in recent times, even computer-controlled players are now impossible to beat. But video games such as Grand Theft Auto are being used to teach AI how to drive a car in an effort to perfect the current self-driving cars on the market. One other skill that is part of living as a society, and that is relevant for aspects like medical procedures is teamwork. Researchers are taking advantage of the level of teamwork necessary to succeed in StarCraft II to teach AI how to play in teams effectively.
Most likely, games played in the hardest mode can easily defeat the average gamer, but what about gambling and poker. Never a machine has been able to take down a professional poker player before until 2018 when a few bots could win over experienced Texas Hold’em players. This trial was done with a robot, named Pluribus, though it can’t yet beat every single poker player. Pluribus has been tested for over 10,000 hands, and it has proven to defeat humans all along. It can take two, three, four, or five players at the same time. Even though professional poker players are not looking to play against machines, this is an essential skill that AI can have that can later be used to better understand the needs of the game.