Exploring Linguistic Adaptations in AI-Mediated Communication: A Sociolinguistic Analysis of Human-AI Interaction

Abstract

Artificial intelligence is increasingly being integrated into our everyday communication practices and its adoption has led to changes in language production, organization and meaning-making in digital communication. While research on communication with artificial intelligence (AI) is increasing, there is a limited sociolinguistic understanding of how people adapt their language to interact with AI. Current studies indicate that AI affects language normalization, discourse structure, and identity construction; yet, research has yet to converge and has primarily focused on technological, rather than fine-grained, linguistic and social meaning making. This paper fills this research gap by providing a sociolinguistic perspective of linguistic adaptations in AI-mediated communication. The study draws on Communication Accommodation Theory, theories of Human-Machine Communication, as well as sociolinguistic theories of variation and identity, which account for language-change processes resulting from communicative adaptations to perceived communicative affordances, including artificial communication systems. In relations of research systems, the study proposes a multimethod approach. Quantitatively, it uses corpus linguistics-based linguistic analyses with natural language processing techniques to investigate lexical richness, grammatical intricacy, tone, discourse markers and statistical tools (including t-tests, ANOVA, regression, and structural equation modeling) to test hypotheses and interpret analyses. The qualitative component includes interviews and discourse analysis techniques such as thematic analysis and critical discourse analysis to understand users' perceptions, negotiations of identity, power, and agency in their interactions with AI. The research is anticipated to theoretically extend sociolinguistic and communication theories to human interactions with AI and methodologically to incorporate quantitative and qualitative analyses. It offers practical implications on inclusive AI system design and AI's role in language usage. The authors suggest future research to empirically verify the proposed model through sequential quantitative, qualitative, and mixed methods design, and to address cross-cultural and longitudinal aspects of AI communication.