From AI Use to Responsible AI Pedagogy: A Framework for Language Teaching in Non-Linguistic Higher Education

Abstract

The rapid diffusion of generative artificial intelligence has created a new pedagogical reality for language teaching in higher education. This article develops a responsible AI pedagogy framework for language teaching in non-linguistic universities, where language courses support not only general communication but also academic literacy, professional vocabulary and discipline-oriented written production. The framework is derived from the qualitative component of a larger mixed-methods case study conducted at the Armenian State University of Economics. The broader study involved a student survey and professional conversations with six lecturers from the Department of Languages; the present article focuses primarily on lecturer-based qualitative evidence and uses student data only as contextual background. Thematic analysis identified five interrelated areas of concern and opportunity: AI as a support tool for linguistic practice, prompt formulation as a form of language thinking, the pedagogical value of comparing student-written and AI-generated texts, risks of dependence and academic dishonesty, and the need to redesign assessment under conditions of widespread AI use. On this basis, the article proposes a five-stage framework consisting of orientation and rule clarification, guided use, critical evaluation of AI output, student-owned language production and reflective AI-use declaration. The article argues that AI should neither be rejected as a threat nor accepted as an unregulated shortcut. Its educational value depends on teacher mediation, transparent rules, task design and the preservation of learner agency. The proposed framework contributes to responsible AI integration in language education by translating ethical principles into concrete pedagogical practice.