The rapid development of artificial intelligence is profoundly reshaping the educational ecosystem. While enhancing educational efficiency, AI also gives rise to ethical risks at multiple levels. From the perspectives of technology ethics, the ethics of responsibility, and educational ethics, this paper systematically analyzes how ethical risks arise in the educational application of artificial intelligence. Based on this analysis, the paper proposes legal and ethical regulatory approaches tailored to various stakeholders. The study emphasizes the need to strengthen multi-dimensional regulatory mechanisms-covering algorithm governance, data security and technology ethics-through collaborative efforts among government, research institutions and educational subjects. It advocates for reconstructing the ecosystem of “human, technology, education” to achieve a harmonious integration of educational values and technological advancement, thereby promoting sustainable and healthy development of AI-powered education.