Until recently, jokes were reserved for humans. Microsoft, In an effort with The New Yorker, has taught a robot to recognize humor. Researchers have fed a robot enough cartoons from The New Yorker to pick out the funny ones. Every month this column shines a light on the technology of our future- this month’s Fireside Chats is on Microsoft’s Artificial Intelligence that has a sense of humor.
Every week Since 2005, The New Yorker publishes a cartoon that’s missing a caption. Readers submit their funniest caption in an attempt to win the contest. The entries are reviewed and the winner’s caption is printed in the following weeks issue. The contest has become very popular and draws about 5000 submissions every week.
Microsoft has taught the robot The New Yorker’s specific brand of humor to reduce the weekly workload of The New Yorker’s staff. Researchers fed an archive of The New Yorker cartoons to a robot in an effort to teach it to detect humor. Microsoft’s robot still has a lot to learn, but all of the human editors top picks appeared in the top 55.8% of the robot’s choices. This alone would cut down the work of a human editor by nearly a half.
Teaching artificial intelligence to understand humor is an important step towards a better experience when interacting with robots, computers or digital assistants like Siri or Cortana. Humor in some form is universally enjoyed- hopefully these robots are hilarious.