КРИТЕРИИ ИСКУССТВА В ИСКУССТВЕННОМ ИНТЕЛЛЕКТЕ

Abstract

The article discusses the development of criteria applicable in art and artificial intelligence. The key points that are currently being created, transformed and already included in the Code of Ethics for Artificial Intelligence are touched upon. The formation of basic provisions and recommendations, involving the country’s leading organisations, is the ethical foundation for today’s understanding of artificial intelligence. The classical approach to art, a description of what was seen or admiration for an author’s work, is given using epithets, metaphors, and figurative expressions. Perception by the classical school, on an emotional level, evokes automatic poetic statements and inferences. Conclusions are based on the accumulated experience of mankind. In AI this experience is selectively invested by a person with the help of some objects (photos, artist’s paintings, drawing techniques). Thus, the experience is transferred by embedding certain knowledge and the neural network can be trained. It is precisely this analogy that the AI developers wanted to achieve; in other words, an analogue of a social structure, where knowledge and accumulated historical experience are passed on from school, which is called the learning process. How close we come to the perception by generally accepted human canons of the results of AI activity in creation, whether we can apply these canons or there should be other criteria and how many of them, is discussed in this article on a specific model of artistic style training.