【摘要】 随着人工智能在教育场景中应用的不断深化，智能教育以其个性化、定制化等独有优势改变了当下的教育格局，但智能教育也面临着算法歧视的潜在风险，如教育结果的单一化、加剧教育不公平、学习结果两极分化、人机交互加深偏见等。同时，在智能教育系统的开发与应用过程中，设计团队存在思维定势、机器学习的偏差、交互式决策的危机等问题使算法歧视的成因更为多样化。而解决对智能教育系统中存在的算法歧视问题，需从促进多主体共同参与，构建多元评价标准，以及通过技术赋能等三方面落实治理策略。具体到技术治理，可根据时间维度划分为三阶段治理，即在事前加强数据歧视检测，在事中建立可解释、可审查的算法优化机制，在事后坚持学生在教育教学活动中的主体性。更多还原

【Abstract】 With the in-depth development of artificial intelligence in education, intelligent education has reshaped the current education with the unique advantages such as personalization and customization. However, intelligent education also faces the potential risks of algorithmic discrimination, such as the simplification of educational outcomes, the intensification of educational inequities, the polarization of learning outcomes, and the deepening of prejudice in human-computer interaction. Meanwhile, in the development and application process of the intelligent education system, the mindset of design team, the bias of machine learning and the crisis of interactive decision-making make the causes of algorithmic discrimination more diverse. In order to solve the problem of algorithmic discrimination in the intelligent education system, it is necessary to implement governance strategies from three aspects: promoting the multi-agent participation, constructing multiple assessment criterion, and accelerating technology enabled educational governance. In terms of technology enabled educational governance, it can be divided into three stages according to the time dimension: strengthening data discrimination detection before the event, establishing the explainable and reviewable algorithm optimization mechanism during the event, and insisting on the subjectivity of students in educational teaching activities after the event.更多还原