Abstract
Artificial intelligence (AI) has become a transformative force in higher education, particularly in language teaching and learning. Tools such as Grammarly and ChatGPT provide learners with personalized feedback, automated grammar and vocabulary correction, and interactive opportunities that enhance engagement and reduce anxiety. These innovations are acknowledged for supporting language proficiency and learner motivation. However, researchers also warn that excessive reliance on AI may lead to a reduction in critical thinking, weaken learner autonomy, and create ethical challenges related to data privacy, plagiarism, and algorithmic bias. A cross-sectional survey was conducted among 180 university and language institute students using the AI-LPAC-E questionnaire, which was designed to assess proficiency, autonomy, critical thinking, and ethical concerns. Findings showed that frequent AI users reported significantly higher proficiency scores than infrequent users (t = 3.21, p = 0.002). Speaking skills improved the most, while writing accuracy was lowest (M = 3.65). However, reliance on AI negatively correlated with critical thinking (r = –0.42, p < 0.01) and autonomy (r = –.37, p < 0.01), whereas autonomy and critical thinking were positively related (r = 0.58, p < 0.01), reflecting their mutual reinforcement. The sample included both university and language institute students as well as teachers, among whom educators expressed greater ethical concerns than learners (t = 4.02, p < 0.001). These results emphasize the need for balanced pedagogical strategies and clear institutional guidelines to ensure responsible AI integration in language education. The findings suggest that while AI tools can effectively enhance proficiency and engagement, educators and policymakers must promote critical thinking and learner autonomy through guided use and ethical training.