No matter how hard we try, humans can't make truly random decisions. Our experiences, beliefs, and mindsets drive our actions, allowing AI to recognize patterns in our behavior and use to them to win simple games.

In this game, the AI learns by making random decisions at the beginning and then adjusting choices to reinforce actions that lead to wins (positive rewards) and discourage those that lead to losses (negative consequences). In other words, like a human child, the AI will make mistakes and learn through experiences. The more rounds you play, the more behavioral data the AI has, and the more accurate it will become at predicting your moves.

Each player in a game must guess what choice the other is going to make. Competitors might be thinking, should I play like I did last turn? Should I copycat my opponent? Agents in call centers also try to guess their customers' true intent, and vice versa; how can I best solve this issue? Are we moving towards resolution? However, the number of behavioral characteristics in play in a complex human to human interaction is vastly greater than in a simple rock paper scissors game. Afiniti's AI considers millions of data points and learns every day to create perfect pairs around the world in mere milliseconds.