Psychological Dependence on AI-Mediated Language Among University Students from Pakistan: A Sociolinguistic Inquiry

This study explores the phenomenon of psychological dependence on AI-mediated language among university students in Pakistan through a sociolinguistic lens. With the rapid integration of AI tools such as ChatGPT, Grammarly, and other language-assistance technologies into academic practices, students increasingly rely on algorithmic support for writing, thinking, and self-expression. While these tools enhance fluency and accuracy, their habitual use raises critical concerns regarding linguistic autonomy, confidence, inner speech, and academic identity. Using a qualitative, phenomenological research design, data were collected from 32 undergraduate and postgraduate students across three Pakistani universities through semi-structured narrative interviews, think-aloud writing protocols, and reflective language diaries. Thematic analysis reveals patterns of cognitive offloading, erosion of self-confidence in unassisted language production, hybrid and fragmented academic identities, and the internalisation of algorithmic norms into students’ inner speech. The findings conceptualise this phenomenon as a form of “algorithmic co-authorship,” where AI functions not merely as a tool but as a non-human socialising agent influencing thought and expression. Situated within the postcolonial and multilingual context of Pakistani higher education, this research contributes to critical sociolinguistics, digital literacy studies, and AI-in-education discourse by highlighting the psychological and sociocultural consequences of AI-mediated communication in the Global South. The study calls for pedagogically informed and ethically grounded approaches to AI integration that promote critical AI literacy, protect linguistic agency, and preserve the human voice in academic knowledge production.