人工智能伦理教育与技术实践的深度融合已成为新工科建设的重要议题。针对传统机器学习实验教学“重技能轻价值”的困境,研究构建“知识–能力–价值”三维目标体系,创新提出“技术解构–社会关联–批判实践”的递进路径。通过双轨案例设计,将数据偏见修正、算法可解释性等思政要素转化为可编程的模型约束条件,结合仿真情境、跨学科协作及开源伦理实践,形成“技术规范与价值判断”协同的教学范式。研究揭示了理工科课程思政的核心机制,即依托技术实践的价值负载特征,在模型调试中实现“伦理认知–价值内化”的转化。为实现教育创新制度化,建议构建校企联动的动态案例库、开发嵌入式伦理评估工具包,并建立涵盖技术性能与伦理敏感度的双重评价体系,推动思政教育从课程改革向系统建构转型,培养兼具技术理性与价值判断力的新型AI人才。The deep integration of AI ethics education with technical practices has become a critical topic in the development of emerging engineering education. Addressing the challenge that traditional machine learning experimental teaching tends to emphasize skills over values, this study constructs a three-dimensional objective system encompassing knowledge, capability, and values. It innovatively proposes a progressive approach through “technical deconstruction-social association-critical practice”. By designing dual-track cases, it transforms ideological and political elements, such as bias correction in data and algorithm explainability, into programmable model constraints. Combining simulated scenarios, interdisciplinary collaboration, and open-source ethical practices, it forms a teaching paradigm where “technical norms and value judgments” are synergized. The research elucidates the core mechanism of ideological and political work in science and technology courses, highlighting the value-laden characteristics of tech
