IBM’s artificial intelligence less convincing than a human

by bold-lichterman

Victorious in chess, IBM’s artificial intelligence, on the other hand, failed to convince an audience during a debate on a social issue organized against a human, said the IT group on Tuesday. At the end of this debate organized in public Monday in San Francisco (west), it is Harish Natarajan, champion of debating competitions, who was declared winner against the computer “Project Debater” stuffed with artificial intelligence, a IBM explained.

The two applicants had had 15 minutes to prepare a discussion on the value of subsidizing preschools: IBM’s program advocated for subsidies, believing it to be an important tool to help the most disadvantaged while his human opponent asserted that this did not address the root causes of poverty and was rather a ” political gift »To the middle classes. Victory depended on their ability to change their mind. Before the debate, 79% of the public thought that nursery schools should be subsidized and 13% thought the opposite. But after the discussion, they were only 62% to be in favor of the subsidies while the proportion of opponents had risen to 30%.

However, IBM felt that there was no question of ” lose or win “But rather to create an artificial intelligence” able to master the infinitely complex and rich world of human language “. It was a new challenge for the American group after the victory of its supercomputer Watson against a great chess player or that in the game of general culture Jeopardy. For Ranit Aharanov, head of the “Project Debater” team, artificial intelligence can help humans make difficult decisions, by “ helping to understand the two sides of a problem ”.