The vast majority of video games incorporate some form of artificial intelligence (AI). From the ghosts in arcade game Pac Man to the bots in first-person shooter Unreal, AI is used to give life to countless characters.
The most basic game SA20 Points Table AI uses simple pathfinding and finite-state machines to model behavior. Pathfinding gets the AI from point A to point B, while state machines permit transition between different behaviors. In more complex video games, the AI may use Monte Carlo tree searches to find ways to overcome an obstacle.[1]
Gameplay-enhancing AI enables NPCs to interact more naturally with players. It also allows them to respond to the player’s actions, which can make them more challenging to defeat or bring a more engaging story to life. The reboot of XCOM, for example, used advanced AI to allow enemies to vary their offensive and defensive abilities based on their own unique strengths.
Procedural Content Generation: AI Building Game Worlds
In more recent years, developers have been able to leverage reinforcement learning techniques to develop even more advanced AI for gaming. Rocket League, for instance, lets players work with robot teammates who are trained through reinforcement learning to perform at blistering speeds during competitive matches.
Other uses of AI in gaming include data mining to discern the ability and emotional state of players, enabling developers to adapt game difficulty and pacing accordingly. Procedural content generation also allows for the automation of repetitive tasks, like creating scenery and music for a game.