Epistemic Modality in English-Medium Artificial Intelligence Research Articles: A Systemic–Functional Perspective

Abstract

Epistemic modality is a crucial tool for describing the speaker's judgment—or request for the judgment of the listener—that has gained broad recognition and been widely investigated in academic writing. This research investigates the use of epistemic modality in 50 English-medium artificial intelligence (AI) research articles (RAs) from a systemic–functional perspective. Our research focuses on the frequency and function of the value, orientation, and polarity of epistemic modality in AI RAs. The results reveal that these AI RA writers tend to use both high- and low-value modality with explicitly objective orientation and positive polarity and that the epistemic modality performs the interpersonal function. This indicates that AI RA writers tend to ensure the objectivity of their results and that these writers adopt varying attitudes and hold differing views toward AI technology and thus have not yet achieved consensus. The study furthers the understanding of scholars’ attitudes toward AI statements in recent years.