Artificial neural networks: New art and transformation of the problem of authorship

This article aims to analyze the transformations of the problem of authorship and a semantic load of art works created through artificial neural networks (ANN) in context of the philosophical tradition. Nowadays, the sale of paintings created by the ANN at art auctions is not a single case. This situation actualizes, on the one side, the issue of a possible institutional recognition of works created by non-humans as art objects, and, on the other - the issue of establishing authorship and a semantic load of such works. The article discusses the conditions for the legitimacy of attributing the ANN practices to art, and analyzes the historical transformation of the problems of authorship and perception of the text (as a sign system) in philosophy. The analysis found out that the research focus had changed during the twentieth century - from the author's totalitarian figure to the problems of language and its internal structures, and then to the reader as a subject entitled with functions of “generating meanings”.