Stylistic Vectors of Using Neural Network Solutions in the Work of a Regional Television Company

Abstract

The paper examines the experience of implementing neural network solutions in text work at the Sverdlovsk Regional Television channel. The editorial team is engaged in the search for a regulation on the use of AI that would not undermine the value-based principles of professional journalism. The alteration in text formation mechanisms under the influence of technological innovations raises pertinent questions regarding the functional-stylistic determinacy of content generation. The terminologically opposed stylistic concepts of M. N. Kozhina, V. G. Kostomarov, as well as researchers of writers’ idiolect, are harmoniously integrated with notions of the distinctiveness of speech and text, as manifested in the functional semantic-stylistic categories and constructive principles (both general and specific). Based on the material of publicly accessible (Yandex services) and proprietary (Media-P.A.U.K.) prompts (requests for content generation by artificial intelligence), the article substantiates the necessity of developing regulations for the use of AI in professional media outlets, demonstrates the demand for textological analysis skills, and justifies the relevance of transferring achievements in functional stylistics to the realm of natural speech. The article proposes a case study of text editing, which has proven effective at the Sverdlovsk Regional Television. The case demonstrates prompts for the initial editing of journalistic text along the stylistic vectors "speaking clearly and accessibly" and "eliminating inappropriate embellishment." The diversity of specific practices of AI usage by editorial teams of television companies brings specialists back to the functional-stylistic characteristics of speech and contrasts the unconscious use of technological advancements with a scalable professional editorial approach.