This study investigates Vietnamese EFL students’ attitudes toward AI-based assessment of English speaking and writing skills and identifies the factors shaping these attitudes. An explanatory sequential mixed-methods design was employed. Quantitative data were collected through a questionnaire administered to 286 English-major students, followed by qualitative data collected through semi-structured interviews with 15 students. The survey data were analyzed using descriptive statistics, while the interview data were analyzed thematically. The findings indicate that students generally hold positive attitudes toward AI-based assessment, particularly regarding its objectivity, efficiency, potential to reduce human bias, and ability to provide timely feedback. However, students also expressed concerns about AI’s ability to fully assess the complex dimensions of speaking and writing proficiency, including communicative intentions, contextual meaning, and creativity. They further emphasized the essential role of human examiners and favored a hybrid approach in which AI supports rather than replaces human judgment in the assessment process. The findings suggest that the use of AI in language assessment should balance technological efficiency with human professional judgment.