AUTHORIAL VOICE AND STANCE IN HUMAN VERSUS AI CO-WRITTEN ACADEMIC TEXTS: A CORPUS STYLISTICS STUDY

Abstract

With the advent of AI writing tools in the academic landscape, the process of creating written content has undergone a significant transformation. The introduction of AI writing tools in academic settings has revolutionized the creation of written content. In this study, the authorial voice and epistemic stance of texts authored by humans and texts co-authored by humans and AI tools (such as ChatGPT and similar LLM tools) is examined. The study utilizes a corpus stylistics method in that the corpus consists of 120 academic texts, which are taken from published journals in the United Kingdom, the United States and Canada. This focus is on indicators of stance, hedging, boosting, self-mention and reader engagement. The results indicate that they found that AI co-written texts have more hedged and neutral language, while the human written texts are more personal, have a wider range of stance expressions, and have a more distinct writing style. The study is based on the Appraisal Theory the main theory used which is based on the work done  by Martin and White (2005). The findings have significant implications for the assessment of academic writing, integrity of authorship, and the future of academic communication in a world where AI tools are increasingly becoming a part of the academic landscape.