Artificial intelligence with 10 billion data could not defeat a person in dispute

Artificial Intelligence Project Debater from IBMfor several years in a row, he has been learning to prove his opinion to people and to win disputes. In June 2018, he opposed Israeli lead Dan Zafir in a dispute over the allocation of subsidies to space exploration and telemedicine and defeated him. The other day, he ran into a more serious opponent - world champion in debate, Harish Natarajan. They met at an IBM conference in San Francisco and talked about the benefits and harms of subsidizing preschool education. The dispute turned out to be very tense and interesting.

The topic of the debate was not known in advance, but announcedjust before the start of the event. The task of artificial intelligence Project Debater was to convince viewers of the benefits of subsidies, and Harish Natarajan opposed this. The participants were given 15 minutes to prepare - during this time, the AI ​​studied 10 billion news items, and Harish sketched a pan speech.

Debates were divided into parts: four-minute speeches, answers to the opponent's arguments, and a closing word lasting two minutes. Project Debater began his presentation with a little joke:

I heard that you hold a world record in the debate against the people, but I believe that you have never argued with the machine. Welcome to the future!

In defense of subsidies, artificial intelligenceannounced that government support is able to protect children from dysfunctional families. In his opinion, helping those in need should be a moral obligation of every person. Natarajan responded to these words by saying that subsidizing does not necessarily mean guaranteed improvements in the quality of education. According to him, such measures may be simple distribution of money to middle class, and people in need may not receive help.

When Natarajan suggested that some childrenthey may not need a kindergarten garden, artificial intelligence reinterpreted his words as "My opponent declared that kindergartens are harmful." Despite such taunts, the champion of the debate presented more specific and convincing arguments than the Project Debater.

The judges were 416 people from the audience. With the help of smartphones, they voted before and after the debate. During his speech, Harish Natarajan reassured 17% of the audience, so he was awarded the victory to him. Artificial intelligence, in turn, was able to bring viewers much more new knowledge than man. According to Harish, the Project Debater is a very strong contender who gives out quite convincing and context-relevant arguments.

What do you think will happen whenartificial intelligence will learn to convince people that they are right? Write your suggestions in the comments, and do not forget to subscribe to our Telegram-chat, where you can discuss other news of science and technology!