The problem of comparative evaluation of paintings created by an artist and generated by a neural network

• Published in: Izvestiâ Saratovskogo universiteta. Novaâ seriâ. Seriâ Filosofiâ. Psihologiâ. Pedagogika (Online) Vol. 23; no. 3; pp. 326 - 330
• Main Author: Panteleev, Alexander F.
• Format: Journal Article
• Language: English
• Published: Saratov State University 22.09.2023
• Subjects: activity; artificial intelligence; digitalization of art; neural network; perception of a picture
• ISSN: 1819-7671; 2542-1948
• DOI: 10.18500/1819-7671-2023-23-3-326-330

Introduction. The role of art in the life of society is to translate humanitarian values. The creator of a truly imaginative work is able to perceive when interacting with the world and embody in his works what may not be in statistics and will probably be relevant for humanity in the future. Theoretical analysis. According to the available data, the pictures aggregated by the neural network and painted by the artist do not differ in subjects. In this paper, an activity-based approach is implemented to assess the differences of such paintings. It is based on the decision of the test subject of the problem associated with the identification of this difference. The difference lies in the method of “translating” the idea of the painting from a symbolic code into a figurative one, objectifying the artist’s attitude to the world, translating humanitarian values in this way. Empirical analysis. In the course of an empirical study, a pairwise comparison of paintings was carried out, as a criterion, the comparison was proposed based on the possible success of the transfer of the artist’s attitude to the world. Conclusion. It was previously stated that the distinction between paintings written by the artist and those generated by the neural network in the conditions of solving a new task for the subject to identify the success of the transfer is problematic.