Algorithmic sovereignty and the future of Armenian: building language tools beyond extractive AI

This conceptual and comparative study examines how artificial intelligence reshapes linguistic sovereignty, using the Armenian language as a revealing case of wider struggles over data, culture, and power and situating these struggles within language policy and planning. Drawing on decolonial AI ethics, Indigenous Data Sovereignty, and critical theories of linguistic justice, it introduces algorithmic sovereignty as the right of communities to govern the digital representation and use of their languages. Through analysis of Armenian’s East–West split, data scarcity, and diaspora context, the study proposes a techno-sovereign framework grounded in community-curated corpora, participatory governance, and dialect-inclusive design. It argues that AI-mediated language infrastructures increasingly function as sites of language planning by shaping linguistic representation, visibility, access, and domains of use, while algorithmic sovereignty foregrounds community authority over these processes. Comparative examples from Māori and African AI initiatives provide context-sensitive comparative lessons rather than universally transferable models. Reclaiming linguistic agency within AI infrastructures, the paper argues, is both an ethical intervention and a contribution to understanding language planning in increasingly AI-mediated linguistic environments.