Idiolect markers in deepfakes as a method of identifying the subject of content in legal practice

the purpose of the article is to develop a method for identifying the author of deepfake content based on idiolect markers for use in legal practice. The relevance of the problem for legal practice and theory is very high: deepfakes create new risks for the reliability of evidence, violate individual rights, therefore methods of their reliable identification are needed, such as the analysis of idiolect markers. The accuracy and operationality of the methodology are important for correct legal decisions in the context of digitalization. The article presents a methodology for developing a multi-level idiolect profile of the subject of speech and an algorithm of actions to establish the probable authorship of the text, necessary for correct legal decisions in the context of digitalization. The methodology shows idiolect markers at phonetic-orthographic, morphological, lexical-semantic, syntactic, and pragmatic levels. The algorithm of actions in legal practice when establishing an idiolect profile is described. The method of multilevel analysis of the idiolect profile of the subject of speech can be used in forensic linguistics to attribute any anonymous texts, establish plagiarism or identify falsifications, the problem in implementing a multilevel analysis of the idiolect profile is the individual variability of linguistic units in the idiolect, an individual language variant. The article substantiates that the development of a methodology for analyzing the idiolect profile requires solving the problem of determining the boundaries of variation in idiological features.

deepfake, idiolect markers, multilevel idiolect profile, idiolect systematization, author's attribution, legal linguistics