Recognizing the artificial: A comparative voice analysis of AI-Generated and L2 undergraduate student-authored academic essays

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00216208%3A11320%2F26%3AVZGBB7J8" target="_blank" >RIV/00216208:11320/26:VZGBB7J8 - isvavai.cz</a>

Výsledek na webu

<a href="https://www.scopus.com/inward/record.uri?eid=2-s2.0-85216954481&doi=10.1016%2Fj.system.2025.103611&partnerID=40&md5=89ebf07bd9360c5cba262c94e5fb137e" target="_blank" >https://www.scopus.com/inward/record.uri?eid=2-s2.0-85216954481&doi=10.1016%2Fj.system.2025.103611&partnerID=40&md5=89ebf07bd9360c5cba262c94e5fb137e</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1016/j.system.2025.103611" target="_blank" >10.1016/j.system.2025.103611</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Recognizing the artificial: A comparative voice analysis of AI-Generated and L2 undergraduate student-authored academic essays

Popis výsledku v původním jazyce

Recent developments in AI content generation software have rendered human-authored texts indiscernible from AI-generated ones. This progression introduced challenges in the academic field, raising concerns over the devaluation of academic integrity. As an inevitable aspect of written texts and an indicator of writing identity, authorial voice is a potential distinguishing factor between both text types. Given this, the researchers examined the differences in the authorial voice of student-written and AI-generated essays using 12 student-written academic essays and 12 AI-generated academic essays. The samples were coded and analyzed using Lehman and Sułkowski's (2020) Voice Analytic Rubric. The findings revealed that Collective (C) voice was the dominant voice in student-written essays whereas AI-generators primarily employed Individual (I) voice. Further comparison and closer analysis showed that: a) I-voice is not the consistent dominant textual identity for AI-generated texts; b) AI-generated text is closer to an expert's writing whereas student-written are closer to a novice's; and c) the writing style of AI-generated texts lean toward predictability. These findings contribute to understanding students' authorial voice construction vis-a-vis the relatively underexplored authorial voice of large language systems. © 2025 Elsevier B.V., All rights reserved.

Název v anglickém jazyce

Recognizing the artificial: A comparative voice analysis of AI-Generated and L2 undergraduate student-authored academic essays

Popis výsledku anglicky

Recent developments in AI content generation software have rendered human-authored texts indiscernible from AI-generated ones. This progression introduced challenges in the academic field, raising concerns over the devaluation of academic integrity. As an inevitable aspect of written texts and an indicator of writing identity, authorial voice is a potential distinguishing factor between both text types. Given this, the researchers examined the differences in the authorial voice of student-written and AI-generated essays using 12 student-written academic essays and 12 AI-generated academic essays. The samples were coded and analyzed using Lehman and Sułkowski's (2020) Voice Analytic Rubric. The findings revealed that Collective (C) voice was the dominant voice in student-written essays whereas AI-generators primarily employed Individual (I) voice. Further comparison and closer analysis showed that: a) I-voice is not the consistent dominant textual identity for AI-generated texts; b) AI-generated text is closer to an expert's writing whereas student-written are closer to a novice's; and c) the writing style of AI-generated texts lean toward predictability. These findings contribute to understanding students' authorial voice construction vis-a-vis the relatively underexplored authorial voice of large language systems. © 2025 Elsevier B.V., All rights reserved.

Druh

J_SC - Článek v periodiku v databázi SCOPUS

CEP obor

—

OECD FORD obor

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Projekt

—

Návaznosti

—

Rok uplatnění

2025

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název periodika

System

ISSN

0346251X

e-ISSN

—

Svazek periodika

130

Číslo periodika v rámci svazku

2025

Stát vydavatele periodika

US - Spojené státy americké

Počet stran výsledku

34

Strana od-do

1-34

Kód UT WoS článku

—

EID výsledku v databázi Scopus

2-s2.0-85216954481