IBM AI Loses Debate to Human Champion

first_imgStay on target In the battle of man versus machine, crushing defeats by DeepMind’s artificially intelligent Go and StarCraft gaming systems tipped the scales in favor of automation.But humankind is making a comeback.On Monday, IBM’s seven-year-old AI debating system Project Debater went head-to-head with a mortal world debate champion Harish Natarajan in a public contest.The pair, moderated by Intelligence Squared host John Donvan, delivered arguments for and against the resolution “We should subsidize preschool.”With just 15 minutes to prepare for the event, Project Debater and Natarajan each delivered a four-minute opening statement, four-minute rebuttal, and two-minute summary.A tall, sleek monolith with a rectangular mouth of blue sound waves, the IBM machine’s knowledge base consists of about 10 billion sentences, collected from newspapers and journals.In a live debate, it searches for short pieces of text in its massive digital library to defend or oppose the motion.“This requires a deep understanding of human language and its infinite nuances and very precise stance identification,” creators Ranit Aharonov and Noam Slonim wrote in a blog post. “[That’s] something that is not always easy for humans and is certainly very difficult for computers.”In a matter of seconds, Project Debater can remove redundant argumentative texts, select the strongest remaining claims and evidence, arrange them by theme, and create a narrative of support or contention.It also listens to and digests the opponent’s responses, building a counterstatement in a “surprisingly charming and human-sounding” way, according to Donvan.Project Debater debuted in June, participating in its first live public event before a small audience. At this week’s IBM Think conference in San Francisco, it faced off against Natarajan in front of a large in-person audience, with many more watching via livestream (video above).The winner was determined by his (Natarajan’s) or her (Project Debater’s) ability to convince the audience of the persuasiveness of their arguments. Results were formulated in a real-time online poll.Before the debate, 79 percent of the audience agreed that preschools should be subsidized, while 13 percent disagreed (8 percent were undecided). After hearing both sides present, 62 percent agreed that preschools should be subsidized, while 30 percent disagreed—clinching Natarajan’s win.Despite the setback, IBM should consider this a win: More than half of participants said that Project Debater better enriched their knowledge about the topic, compared to only 20 percent for Natarajan.“Ultimately, what we saw was that the interaction of man and machine could be enriching for both,” Slonim said in a statement following the debate. “It’s not a question of one being better than the other, but about AI and humans working together.”More on Geek.com:IBM and McCormick Tap AI to ‘Spice Up’ Food SeasoningsIBM Releases Diverse Dataset For Fairer Facial RecognitionIBM Designs AI-Enabled Fingernail Sensor to Track Diseases McDonald’s Plans to Serve AI Voice Technology at Drive ThruCIMON Returns to Earth After 14 Months on ISS last_img

Leave a Reply

Your email address will not be published. Required fields are marked *