Poorly trained artificial intelligence (AI) can pose a real danger, says Inna Kramarenko, Candidate of Economic Sciences, Associate Professor of the Department of Mathematical Methods in Economics and Management of the SUM. In an interview with Lenta.ru, she explained what his main danger was.

“Artificial intelligence is not matter that appeared from another planet and is able to independently cognize the world around us, without relying on existing cause-and-effect relationships, rules and laws,” the specialist noted. Knowledge of the world of AI, according to her, is carried out on the basis of billions of provided examples. Therefore, it is important what material is provided to artificial intelligence for learning.

The danger of AI lies in the fact that examples for learning can be presented with a deliberate or random error, in which case the AI will produce an inadequate result.

Inna Kramarenkoeconomist

In addition, it is not only not expedient, but also dangerous to replace a person with AI in some areas, she is convinced. So, according to the interlocutor of Lenta.ru, the implemented AI in the form of bots in social networks and call centers is able to influence the consciousness and have a psychological impact on a person, is able to participate in the formation of his worldview.

Without a doubt, artificial intelligence can replace a person when performing the same type of routine operations, and it should help a person cope with the tasks of searching, identifying, modeling in endlessly changing information flows. But AI is also capable of writing letters, broadcasting, replacing a person in those areas in which it is unethical to do this, the specialist drew attention.

“Many countries are already thinking at the legislative level about limiting the areas for the introduction of artificial intelligence. Of course, a “golden mean” should be found in the use of artificial intelligence, and AI should always be used for the benefit of the development of society and humanity,” Kramarenko summed up.

Related materials:People chipping, DNA editing and an alternative to antibiotics.How has global medicine changed in 2022?January 5, 2023"One hundred percent success"How artificial intelligence helps men and women conquer a person at first meetingJanuary 11, 2023

Earlier, Microsoft co-founder Bill Gates said that neural network systems like the ChatGPT bot will change the world.

In early February, it became known that a team of specialists from the University of Waterloo had developed a neural network that was taught to predict the effectiveness of chemotherapy for breast cancer.