Slashdot Mirror

← Back to Stories (view on slashdot.org)

IBM's AI Loses To a Human Debater (cnet.com)

Posted by ryuzaki0 on from the how-about-that dept.

The subject under debate was whether the government should subsidize preschools. But the real question was whether a machine called IBM Debater could out-argue a top-ranked human debater. The answer, on Monday night, was no. CNET: Harish Natarajan, the grand finalist at the 2016 World Debating Championships, swayed more among an audience of hundreds toward his point of view than the AI-powered IBM Debater did toward its. Humans, at least those equipped with with degrees from Oxford and Cambridge universities, can still prevail when it comes to the subtleties of knowledge, persuasion and argument. It wasn't a momentous headline victory like we saw when IBM's Deep Blue computers beat the best human chess player in 1997 or Google's AlphaGo vanquish the world's best human players of the ancient game of Go in 2017. But IBM still showed that artificial intelligence can be useful in situations where there's ambiguity and debate, not just a simple score to judge who won a game. "What really struck me is the potential value of IBM Debater when [combined] with a human being," Natarajan said after the debate. IBM's AI was able to dig through mountains of information and offer useful context for that knowledge, he said.

5 of 95 comments (clear)

  1. Possible bias? by sheramil · · Score: 3, Interesting

    The victory was decided by the audience, who knew they were listening to a machine, and they may have been biased against it for that reason. A couple of people may even have had a bias for the machine's argument for that reason.

    1. Re:Possible bias? by thereddaikon · · Score: 3, Interesting

      Seems to me the best way to do this is to have an actor on stage representing the AI and delivering its lines. That way the audience isn't sure who is the real debater and who is the AI. Almost making it "Turing Blind" in a way. Then again, debating is a sort of indirect Turing test. I think there is a statistically significant portion of the population who would never be swayed by the rhetoric of a robot, no matter how sensible it was. But if they didn't know it was a computer, they would be more receptive.

    2. Re:Possible bias? by stevegee58 · · Score: 3, Insightful

      Ben Shapiro and Jordan Peterson would have both "lost" in the judgement of any Left Coast audience.

  2. Re:Debate? by Nidi62 · · Score: 3, Insightful

    When dealing with people, numbers are not the end all, be all. There are times when the quantitatively correct solution may not necessarily be the qualitatively correct one. Say for example there is a disease that can be treated through regular yet painful treatments at the cost of $1 million. There is a cure for the disease, with a one-time application that costs $1.5 million. Quantitatively the treatment course is the best option as it it cheaper. However, qualitatively, the cure is the best option as it reduces suffering.

    For a more real world example, let's looks at the Titanic sinking and the classic "women and children first". From a purely quantitative point of view, it would have more optimal to prioritize men and women of economical or child-bearing productive age as they have the most benefit to society, then the children, and finally the elderly. However, no one would accept that solution as the most optimal one, neither then nor now.

    --
    The only thing necessary for evil to triumph is for it to be pitted against a slightly greater evil
  3. Re: So universities determine intelligence? by steveb3210 · · Score: 3, Interesting

    Plenty of state schools and the like also have world class professors.... At UMass I took AI with Andrew Barto who's co-authored the go-to text book for machine learning.

    I took discrete math with Neil Immerman who proved NL=CoNL.

    I took Abstract Algebra with Arunas Rudvalis who discovered one of the finite simple groups.

    Not exactly intellectual slouches!