【目的/意义】 随着AI技术全面融入社会生活，AI素养已成为智能时代公众必备的基础性能力。通过系统梳理国内外AI素养测评实证研究的进展，为后续相关研究提供参考和借鉴。 【方法/过程】 遵循系统综述法，共筛选69篇国内外AI素养测评的实证研究作为分析样本，通过遍历文章内容的方式，统计样本文献的描述性特征，并从测评对象、测评维度、测评工具的开发与验证、影响因素及研究局限等方面进行系统梳理。 【结果/结论】 现有AI素养测评的对象主要包括社会公众、学生、教育工作者；测评的核心维度包括认知、应用、评估、伦理；测评流程涵盖样本选取和测评工具的开发与验证；AI素养水平会受到性别、年龄、国籍、教育背景等变量的影响；现有研究存在测评工具、测评样本、测评方式等方面的局限性，有待在未来研究中进一步完善。

[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.