Asymmetrical pluralism and the normative power of artificial intelligence in Asian Englishes

When a speaker of Singapore English types ‘Can or not?’ into an AI writing assistant and receives ‘Is it possible?’ in return, something more than correction has occurred: a normative judgment has been rendered, algorithmically and at scale. This article examines how artificial intelligence (AI) systems reshape normative power in Asian Englishes through mechanisms of differential audibility. Focusing on Outer and Expanding Circle Englishes, it argues that AI operates as a normative infrastructure mediating ideologies of correctness, fluency, and legitimacy beyond traditional institutional gatekeeping. Systematic analysis reveals that AI privileges globally dominant varieties while filtering features of Asian Englishes into statistical marginality. This produces asymmetrical pluralism: multiple Englishes are present in AI-mediated interaction but are unevenly authorized in terms of visibility, normative weight, and legitimacy. Despite Asia’s substantial AI engagement, participation does not translate into proportional normative influence. The article calls for greater attention to audibility, norm circulation, and inequality in AI-mediated contexts.