COLONIAL LEGACY TO ALGORITHMIC GOVERNANCE: LANGUAGE POLICY, POWER, AND AI IN PAKISTAN (1947–2025)

Abstract

This paper has analyzed the history of language policy in Pakistan since 1947 to 2025. The aim of the research was to examine the ways language policy produces and reinforces power, ideology, and linguistic inequality between English, Urdu, and regional languages and explore the impact of the emergence of artificial intelligence (AI) as a new player in language policy. A qualitative interpretive research design was used in the study. It was based on interdisciplinary theoretical framework that combines Language Policy and Planning (LPP), Critical Theory, and Algorithmic Governance. The research relied on the secondary sources of data, such as documents on language policy, provisions of the constitution, education policy, AI policy frameworks, and scholarly literature. The data were collected using purposive sampling. Historical analysis, critical policy analysis and interpretive evaluation were the methods of analyzing data. The results indicate that the language policy in Pakistan has changed in terms of its form, but not structure with the hierarchical system of language policy where English is dominant, Urdu is symbolic and regional languages are excluded. An analysis of history and the constitution indicates that there has always been a policy-practice gap in which Urdu is advanced and English remains in power to govern, teach and learn, and socio-economic mobility. Education policies and institutional practices give other pieces of evidence that prove that English is being used as linguistic capital where it strengthens the power of the elite and social inequalities. At the modern stage, AI has become a non-state language policy player, whereby the data availability and technological infrastructure determine the representation of language. Algorithms have a strong preference in English, with partial Urdu support and no support of regional languages like Punjabi, which is the result of limited datasets and technical limitations, and results in algorithmic inequality (Cristaldi, 2025; Sharif et al., 2024; Butt et al., 2025).  The study proves that linguistic inequality is not only traditionally passed on but also digitally replicated, and the policies of inclusive language and AI should be introduced.