This study presents learner agency as the central principle for ensuring successful artificial intelligence (AI) integration in English language teaching and learning. While AI-enabled tools offer a wide range of possibilities for personalized feedback, individualized language training, and access to extensive learning resources, their educational value depends on the extent to which they strengthen learners’ autonomy rather than foster dependence on technology. Drawing on theoretical models of human agency, self-regulated learning, learner autonomy, sociocultural learning, and second language acquisition, the study introduces the AI-Supported Learner Agency Framework for English Language Learning. Within this framework, AI functions as a mediating resource intended to enhance learners’ cognitive, linguistic, and metacognitive abilities rather than replace them. The study emphasizes inquiry-based learning, AI literacy, teacher-scaffolded guidance, feedback literacy, authentic assessment, and ethical oversight. Particular attention is given to English language learning contexts where learners increasingly rely on generative AI to compose essays, translate texts into other languages, develop vocabulary, practice pronunciation, and communicate verbally. More specifically, the study argues that the central educational question is no longer “whether learners use AI,” but “how do learners use it”: as passive recipients of machine-generated answers or as independent and critical thinkers in their engagement with AI.