“Generative artificial intelligence and qualitative accounting research: On the road to epistemic untrustworthiness?”

Department : Accounting and Management Control

Speaker : By Professor Yves GENDRON from University of Laval

Room : V Ramanantsoa

Abstract

This paper examines how the emergence of generative artificial intelligence (genAI) challenges the trustworthiness of qualitative accounting research (QAR), which traditionally hinges on the human dimension. While researchers appear increasingly attracted to genAI’s potential to enhance productivity, its use may disrupt the social and epistemic processes through which qualitative knowledge is produced and vetted. To address these challenges, we draw on Lincoln and Guba’s (1985) criteria for establishing trustworthy qualitative research and Oreskes’ (2019) pillars of trust in science. Specifically, we seek to address the following question: What are the potential implications of using genAI for the trustworthiness of QAR? To this end, we conducted an empirical examination of various genAI tools to illustrate their potential (mis)use in QAR. Our findings suggest that genAI lacks embodied experience and sensitivity to context, resulting in decontextualized interpretations. Our analyses also indicate the ease with which genAI can fabricate qualitative data, creating a previously unthinkable opportunity for misconduct in qualitative research. Ultimately, our examination suggests that the use of genAI could gradually erode researchers’ essential skills. To mitigate these risks, we propose concrete initiatives that advocate for a multi-actor approach, in which accounting scholars proactively address the use of genAI in QAR. In doing so, we reconceptualize trustworthiness as a professional and institutional ethical obligation at the heart of accounting knowledge production, rather than simply a methodological standard.