Zhou Meixia

Guangzhou Huashang College

Abstract：

correlations. The findings are significant for adolescents, as they offer opportunities to predict and explore career development.The 2022 version of the new curriculum standards has established core competencies as the core goal of education, forcing the traditional end-of-term assessment model based solely on test scores to undergo innovation. The traditional score-based assessment focuses on the identification of knowledge outcomes and is unable to conduct quantitative diagnosis of high-order competencies such as critical thinking, problem-solving, and cooperative inquiry; artificial intelligence technologies such as big data, natural language processing, multimodal perception, and knowledge graphs provide a feasible technical path for the transition from a single-score determination to a comprehensive ability diagnosis. This paper distinguishes the essential differences between score-based assessment and competency-oriented ability assessment, explains the internal logic of the transformation of the assessment paradigm driven by artificial intelligence, and outlines the four major practical challenges currently faced in the transformation implementation, namely, insufficient indicator adaptation, algorithm black boxes, human-machine imbalance, and data ethics. It proposes implementation paths from four dimensions: reconstructing competency evaluation indicators, establishing a human-machine collaborative evaluation mechanism, optimizing algorithm reliability and validity, and improving data governance and teacher support, to provide theoretical references and practical solutions for the digital reform of educational evaluation.

Key Words：

new curriculum standards; core competencies; Artificial Intelligence; score-based assessment; "Ability Assessment"; transformation of evaluation paradigm