During a game, Watson had access to 200 million pages of content, including the full text of Wikipedia.[4] Sources of information for Watson included encyclopedias, dictionaries, thesauri, news articles, and books.[5] It was not connected to the internet,[6] meaning that it had to use what was in its system to answer clues. For each clue, Watson's three most likely responses were displayed on the television screen along with its amount of confidence in those answers. If its confidence in a response was high enough, it would ring in to give the response.
After Jeopardy!, Watson continued to be part of IBM's research in artificial intelligence. On new problems such as medical records and genetics, Watson technology was unable to compete with new methods such as deep learning.[7]