[Purpose/Significance] As AI technologies become deeply embedded in social life, AI literacy has emerged as a fundamental competency for the public in the intelligent era. This study aims to systematically review the progress of empirical research on AI literacy assessment worldwide, providing reference and insights for future studies. [Method/Process] Following the principles of systematic review, this study selected 69 empirical studies on AI literacy assessment as the analytical sample. Through a comprehensive examination of the included literature, descriptive characteristics were summarized, and key elements of AI literacy assessment—including assessment targets, assessment dimensions, development and validation of assessment tools, influencing factors, and research limitations—were systematically analyzed. [Result/Conclusion] Existing research shows that AI literacy assessments primarily target the general public, students, and educators. Core assessment dimensions commonly include cognition, application, evaluation, and ethics. The assessment process generally involves sample selection and the development and validation of assessment tools. Empirical findings indicate that AI literacy levels are influenced by variables such as gender, age, nationality, and educational background. Current studies still exhibit limitations regarding assessment tools, sample characteristics, and assessment methods, suggesting the need for further refinement in future research.