AUTHORSHIP ANALYSIS IN FORENSIC LINGUISTICS IN THE AGE OF ARTIFICIAL INTELLIGENCE Cover Image

YAPAY ZEKÂ ÇAĞINDA ADLİ DİLBİLİMDE YAZAR ANALİZİ
AUTHORSHIP ANALYSIS IN FORENSIC LINGUISTICS IN THE AGE OF ARTIFICIAL INTELLIGENCE

Author(s): Hülya Kocagül Yüzer
Subject(s): Morphology, Lexis, Semantics, Psycholinguistics, Sociolinguistics, ICT Information and Communications Technologies
Published by: Artsürem Bilim Sanat Danışmanlık A.Ş (Artsürem Yayıncılık)
Keywords: Forensic linguistics; Authorship analysis; Idiolect; AI Authorship detection; Artificial idiolect;
Summary/Abstract: This study aims to propose a framework for examining how the text-generation capacities of large language models (LLMs) are transforming forensic linguistic authorship analysis. In this context, the conceptual foundations of authorship are interrogated, while its methodological transformation and legal implications are discussed. The theoretical trajectory extending from Barthes’s (1967) critique of authorial authority to Foucault’s (1969) concept of the author function, and further to Love’s (2ii2) functional model of authorship, has evolved into a new field of tension in the age of artificial intelligence (AI). As emphasized by Sousa-Silva (2i24), since AI-generated texts lack individual authorial characteristics, forensic linguistic methods need to shift from surface-level stylistic features to deeper linguistic properties (pp. 164–165). The central argument of this study is that the traditional assumptions of forensic linguistics—namely intra-author consistency (the expectation that the same author exhibits similar linguistic features across different texts) and inter-author distinctiveness (the assumption that different authors display distinguishable patterns) (Grant, 2i22)—may lose their unconditional validity in the context of AI-mediated text production. This situation necessitates a re-evaluation of existing conceptual frameworks. As a contribution, the study proposes a reinterpretation of Love’s (2ii2) functional model of authorship in the context of AI, highlighting the distinction between artificial idiolect and human idiolect (Coulthard, 2ii4). From a methodological perspective, the importance of the multi-layered analysis advocated by Nini’s (2i23) Theory of Linguistic Individuality is underscored, while, in legal terms, the applicability of the Daubert criteria in the age of AI is critically examined.

  • Page Range: 212-243
  • Page Count: 21
  • Publication Year: 2025
  • Language: Turkish
Toggle Accessibility Mode