Recent advancements in generative artificial intelligence (AI) tools have increasingly blurred the boundary between student-authored and AI-generated academic texts, raising urgent concerns about the erosion of academic integrity, creativity, and authenticity in student writing. While the existing body of literature has primarily focused on identifying stylistic differences to detect AI-generated content, this review adopts a distinct and underexplored perspective. This thematic review aims to examine how the use of AI writing tools, particularly in second language contexts, may influence students’ authorial voice and linguistic fingerprint. The primary research questions guiding this study are: What are the key stylistic and linguistic differences between AI-generated and human-authored academic texts, and how do these differences affect the authorial voice and linguistic fingerprint in academic writing? This review concludes by highlighting key pedagogical implications for fostering authentic expression and preserving individual identity in academic writing. Drawing on 18 scholarly papers published between 2023 and 2025, this thematic literature review synthesizes up-to-date empirical findings across three core dimensions: (1) stylistic and linguistic features, (2) authorial voice and linguistic fingerprint, and (3) pedagogical implications for writing instruction. The evidence reveals that although AI-generated texts tend to be grammatically polished and structurally coherent, they often lack the emotional nuance, subtlety, rhetorical richness, and distinct linguistic markers that characterize human-authored writing, which may threaten the linguistic fingerprints of individual writers. This review highlights the urgent need for pedagogical strategies that not only integrate AI tools effectively into academic writing instruction but also actively protect students’ individuality, authorial voice, and creative expression. This is particularly relevant in English as a second language (ESL) university settings, where the majority of the synthesized studies were conducted.
Jokha Al Hosni (Mon,) studied this question.