Development of ethical evaluation indicators for educational artificial intelligence based on the “Life-Practice” pedagogy

The rapid advancement of artificial intelligence has ushered in a new era of educational artificial intelligence (EAI), bringing both immense potential and significant ethical challenges. To promote responsible development and effectively mitigate potential ethical risks, there is an urgent need to establish clear evaluative criteria for ethical practices involving EAI. However, current generic frameworks often prioritize technical safety over pedagogical generativity. This study addresses this imperative by developing an ethical evaluation model and corresponding system of indicators for EAI. Guided by the core principles of Professor Ye Lan’s “Life-Practice” Pedagogy, this study developed a set of ethical evaluation indicators across two critical perspectives: design and usage. These indicators were refined through two rounds of Delphi surveys, and their importance was quantitatively determined using the Analytic Hierarchy Process (AHP). The resulting ethical evaluation framework comprises two perspectives, encompassing 8 first-level indicators and 32 s-level indicators. These indicators offer theoretically informed guidance that requires further empirical validation, thereby potentially supporting the ethically informed development and responsible integration into educational contexts.