IBM is in the final stages of developing a computer program to battle human contestants on the game show Jeopardy!—a move that, if successful, could bring artificial intelligence a big step forward, the New York Times reports. To successfully compete, the computer (nicknamed Watson, after IBM founder Thomas J. Watson Sr.) will have to be able to make subtle comparisons and interpretations, understand puns and double entendres, and interact with humans at rapid speed. Researchers are trying to develop a new type of software that can “understand” and respond to human questions, representing an effort by IBM to prove that research can be spurred by picking “grand challenges” like the famous chess battle (IBM was also responsible for the chess program that ran on supercomputer Deep Blue, beating champion Garry Kasparov in a 1997 match). Under rules negotiated by the show’s producers and the research team, the computer will receive questions as electronic text, and respond with a synthesized voice. Its answers will be based on text it has “read” before the show. In a recent demonstration at the IBM lab, Watson beat human competitors to the question, “Bordered by Syria and Israel, this small country is only 135 miles long and 35 miles wide,” by answering, “What is Lebanon?” However, the program also made errors, for example identifying “sheet” as a fruit.