Creativity of Neural Networks in Modern Scientific Research

Abstract

The topic linking scientific achievements with breakthroughs in AI has moved into the focus of readers' interest over the past decade. Thematic issues of high-ranking domestic and foreign journals are devoted to its discussion. The authors of Information Society, following the chosen editorial policy, contribute to the comprehension of issues related to this problematic. The main idea of the article lies in the author's aspiration to answer several questions related to the spread of the influence of generative AI's enormous potential on the entire course of contemporary scientific research. The author's argumentation regarding the creativity of neural networks in scientific research is supported by concrete examples of achievements in this field. Along with this, the article demonstrates that the achievements of twentieth-century philosophy of science are by no means taken outside the boundaries of AI-related themes; on the contrary, they contribute to its deeper comprehension in an evolutionary key. In the concluding part of the article, the question is raised of filling the concept of "scientific understanding" with new meanings under conditions of the growing creative potential of neural networks.