The tendency to "animate machines", which arose in the world of scientists at the very beginning of this century, not only does not lose its popularity, but, on the contrary, increases it. One gets the impression that scientists around the world are striving to make the plot of the famous film "The Terminator" come true.
Smart robots are already able to do a lot of things that were previously available only to humans. For example, they are capable of generating good texts and maintaining a conversation. Many are amused by the opportunity to talk to the "voice assistant", which is now available in almost every browser. But how many people think that in fact a robot has replaced it with friendly communication with real people?
Meanwhile, scientists are very actively continuing to engage in various research and create more and more similar to the human mind samples of artificial intelligence. The theme of the next work in this area is humor.
That is, it is quite possible that in the near future jokes will cease to be a human prerogative. And the artists of the spoken genre will be replaced by a cute robot with the skill of stand-up.
Where are they trying to create robot humorists?
Scientists at two American universities - at Stanford and at Purdue - are trying to teach artificial intelligence not only to reproduce jokes, but also to understand the very nature of laughter. Interestingly enough, the research of specialists does not overlap, although they refer to the same issue.
However, a group of scientists from Stanford University were unable to achieve success in their experiments. Californian experts announced that they managed to "teach" artificial intelligence to make puns, however, the computer mind cannot joke "correctly" yet. Scientists do not give up the hope that the programs necessary for artificial intelligence for learning will nevertheless be written, and the machine will be able not only to understand the nature of humor, but also to make someone laugh.
Promotional video:
But their Indiana colleagues at Purdue Community Life Science Research University have been slow to announce their fiasco. The work is supervised by Julia Raise, Associate Professor of the Department of Computer and Information Technologies. In her recent interview with a university newspaper, she said that "teaching" a computer to joke can only be through finding an opportunity for an accurate assessment and correct interpretation of human intonations, emotions, and speeches.
In other words, in the work of scientists from Indiana, "humor" acts as a kind of test, litmus, on which it is easiest to track the capabilities of artificial intelligence in understanding the duality of human nature.
What is the essence of the work of scientists?
The essence of the experiments conducted by the Americans is to expand the capabilities of the computer mind. That is, at the moment, cars can be very much, but still lag behind people.
The human brain is able to process and interpret information in two different ways - direct and allegorical. The human mind is also capable of building logical chains, of thinking in both versions, and a person can also combine them. At the moment, the computer mind is only given a direct understanding of information, allegory is not available to it.
That is, voice assistants in browsers will not understand sarcasm, will not notice the subtext, will not take hints, and even more so, will not be able to correctly interpret stable phraseological units. But, if the experiments of American scientists are successful, then the allegory will become available to artificial intelligence.
Angelica Braldi