Generative artificial intelligence (GenAI) tools such as ChatGPT and Grammarly have rapidly entered English as a Foreign Language (EFL) education, offering new opportunities for feedback and learning autonomy while raising ethical concerns. This exploratory study examined Vietnamese translation stu-dents' perceptions of GenAI through the combined lenses of the Technology Acceptance Model (TAM) and Self-Determination Theory (SDT). A survey of 44 undergraduates measured self-report ratings on perceived usefulness (PU), perceived ease of use (PEOU), autonomy, competence, and ethical use using a 32-item Likert-scale questionnaire. Results revealed uniformly high self-perceptions across all con-structs, indicating strong subjective acceptance and confidence. Correlational analyses showed signifi-cant positive relationships among all variables, while regression analyses demonstrated that PU signif-icantly predicted competence and autonomy, and competence alone predicted ethical use. These find-ings suggest that students' ethical engagement with GenAI stems primarily from their perceived com-petence rather than ease or usefulness alone. The study highlights a developmental shift from mere AI acceptance toward ethical empowerment, emphasizing competence-based ethics and autonomy-sup-portive pedagogy. Integrating TAM and SDT provides a holistic framework to understand how cogni-tive beliefs and motivational needs jointly shape responsible AI literacy in EFL education. Pedagogical implications include designing guided AI literacy tasks that transform technological utility into ethical capability.