Ethically-oriented agile governance for generative artificial intelligence represents the optimal governance model for responding swiftly and dynamically to ethical risks associated with GenAI. It provides theoretical underpinnings for establishing a systematic ethical governance framework across the entire GenAI ecosystem. Presently, this governance model is confronted with an inevitable transition from theoretical exploration to practical implementation pathways. Accordingly, the practical implementation of ethically-oriented agile governance should take into account three core dimensions: first, establishing ethical consensus principles and scenario-specific standardized guidelines as governance instruments; second, embedding data ethics, algorithmic ethics, and review-evaluation mechanisms throughout the entire lifecycle of intelligent systems during the governance intervention phase; and third, facilitating constructive interaction among governance entities at the macro, meso, and micro levels. In line with the principles of scientific and technological innovation, a robust science and technology ethics framework that is compatible with China's national context should be established.
Yunxin Zhao (Thu,) studied this question.