An AI named Libratus, beat several players at poker winning  1.7 million in fake money, this is concerning for the public because it allows people to see how intelligent AI robots can be. People are astonished by the how intelligent the AI was however but on the contrary, it also makes wonder if the AI robot is too smart. It is truly amazing seeing how far technology has come for computers to be able to beat the people that created them at a game. In 1997 when Deep blue beat Gary Kasparov in chess the world was astonished. To an AI it does not realize what it is doing, it only computes what move will have the best outcome then proceeds with that play. Libratus’s success came from three components. The first component was known as reinforcement learning, it learned how to wage bets against human components by playing games against itself. The second component was called end-game resolver, this allowed Libratus to learn from the games it was playing. The third was a component that prevented its competitors from learning its patterns forcing him to change is game style every time it played.

Read More: https://www.theguardian.com/technology/2017/feb/05/artificial-intelligence-ethics-poker-libratus-texas-holdem-ai-deepstack