李飞飞最新访谈：AI咋能代替人呢？

TL;DR - Fei-Fei Li argues that AI should augment rather than replace people, with human oversight, broad public participation, and domain-specific safeguards guiding adoption. The interview matters because it connects AI’s technical limitations—statistical learning, data dependence, and lack of subjective experience—to practical governance and deployment choices.

• AI still learns largely from statistical patterns in massive datasets, unlike humans’ data-efficient, contextual learning.
• Healthcare illustrates the best near-term model: AI can support diagnosis, research, and robotic surgery, but sparse data and patient variability limit autonomous operation.
• AI cannot genuinely experience empathy, emotion, or private subjective memories; human creativity and judgment therefore remain central.
• Li advocates participatory governance, ethics education, and strong safety oversight for high-risk applications rather than leaving decisions solely to technology companies.