【作 者】：
阴祖宝， 容中逵
【关 键 词 】：
人工智能； 教育研究； 算法； 算法风险
【栏 目】：
理论探讨
【中文摘要】：
在教育数字化转型背景下，基于算法的教育研究已呈现出蓬勃发展之势，正推动着教育研究由以个体数据为主的实证分析向以海量关系数据为主的智能预测转型。基于算法的教育研究具有独特的运行逻辑，表现为指向计算主义的研究本体论规定，指向算法表征的研究认识论预设，指向数据驱动的研究方法论实践，指向模型预测的研究价值论诉求。但基于算法的教育研究也面临量化一切的算法崇拜、人之主体性的僭越、唯数据论的方法圭臬以及教育研究的非正义性等潜在风险。为防范和化解算法风险需要建立相应的规制机制，具体包括嵌入基于人文主义的研究立场，构建主体间性的“人—算”协同关系，推进基于数据效能的算法公开，重塑基于教育逻辑的算法正义等。
【英文摘要】：
In the context of digital transformation of education, algorithm-based educational research has shown a booming trend and is driving the transformation of educational research from empirical analysis based on individual data to intelligent prediction based on massive relational data. Algorithm-based educational research has a unique operating logic, which is manifested as the research ontological prescriptions of computationalism, the research epistemological presuppositions of algorithmic representations, the data-driven research methodological practices, and the research axiological appeal of model prediction. However, algorithm-based educational research also faces potential risks such as the cult of the algorithm that quantifies everything, the arrogation of individual subjectivity, the principle of data-only method and the injustice of educational research. In order to prevent and resolve algorithmic risks, it is necessary to establish corresponding regulatory mechanisms, including adhering to a humanistic research standpoint, constructing an intersubjective "human-algorithm" cooperative relationship, promoting an algorithm openness based on data efficiency, and reshaping the algorithmic justice based on the logic of education.