Artificial intelligence machines have beat the world’s greatest Go, Chess, Jeopardy, and StarCraft II (usually) players, but they’re still not quite on the same level as one of the world’s top debaters—Harish Natarajan, a grand finalist in 2016’s World Debating Championships and the 2012 European Debate winner. He beat IBM’s six-year-old debating computer, known as Miss Debater.
Neither the computer nor Natarajan were informed of the debate subject—whether or not preschool should be subsidized—until 15 minutes before the competition began.
Each competitor had to present a four-minute opening statement, a four-minute rebuttal, and a two-minute closing argument. Project Debater, which isn’t connected to the internet, comes up with statements using data from the 10 billion sentences stored in its database, which are taken from newspapers, magazine articles, and academic journals. It presents these in what’s described as a mostly monotonous, female voice.
The audience, which was made up of journalists, tech industry insiders, top debaters from Bay Area schools, and software engineers, had to vote for one side at the start of the debate and once again when proceedings were over. Natarajan won by changing more people’s minds.
Going into the debate, Natarajan believed he would have an advantage. “I imagine at this stage a human would still find it easier to construct logical arguments than a machine would in a way which is reasonably convincing to a human audience,” he said.
While Miss Debater isn’t quite there yet, it likely won’t be long before debating is another area where AI has us humans beat.