IBM’s AI system that won the 2011 Jeopardy! tournament was called Watson.
Watson competed on the television quiz show Jeopardy! in February 2011 against former champions Ken Jennings and Brad Rutter. It won the two-game match and earned a $1 million prize, while Jennings finished second and Rutter third.
The system combined natural-language processing, information retrieval, machine learning, and many answer-generation and ranking techniques. Jeopardy! questions often use wordplay, indirect clues, and broad subject knowledge, so Watson had to analyze the clue, generate candidate answers, and estimate confidence before responding.
Watson did not function like a simple web-search box. Its performance came from processing large stores of structured and unstructured information and selecting answers rapidly. IBM later used the Watson name for a broader collection of commercial AI products, which can cause confusion with the original quiz-show system.