Development of The concept of social artifi cial intelligence

Abstract

Research in the field of artifi cial intelligence (AI) has revealed a number
of conceptual shortcomings and risks of the developed cognitive models and
AI algorithms when applied in various fields of activity. This, in particular, is
the opacity, non-verifi ability of the logic of reasoning and conclusions of AI,
not friendliness towards a person, and, as a result, distrust of a person in the
results of AI work. Distrust between artifi cial and natural intelligence leads
to a decrease in the eff ectiveness of both AI and its interaction with human
intelligence. As a way to overcome this crisis of misunderstanding, the concept
of artifi cial social intelligence (ASI) was proposed, which includes not only
cognitive aspects of intelligence, but also social (together with emotional ones).
The article sums up the results of thirty years of development of the concept
of ASI, highlights the current problems of modeling and studying ASI.