Song-Chun Zhu, a highly respected AI scientist, left his long-standing academic career in the U.S. in 2020 after nearly three decades to take up leadership roles in China. Born in rural China during the Cultural Revolution and raised amid hardship, Zhu pursued his PhD at Harvard and later led a prolific AI research lab at UCLA. Over time, he grew uneasy with the prevailing AI trends focused on deep learning and large datasets.
His intellectual philosophy emphasizes what he calls a “small data, big task” paradigm—AI systems that can reason, understand physical and social context, and generalize from less information, rather than relying strictly on massive data and computation. Zhu believes this approach better reflects human cognition and could be more effective in developing artificial general intelligence (AGI).
Political pressures, strained relations between the U.S. and China, and increasing scrutiny of Chinese researchers played a role in complicating his situation. In China, Zhu now leads the Beijing Institute for General Artificial Intelligence (BigAI) and holds academic appointments. Though his ideas diverge from mainstream AI practices, he insists that his motivations are scientific, not political.
His story illustrates the intersections among scientific ambition, intellectual freedom, and geopolitics in today’s AI landscape. Zhu argues that access to resources and the ability to pursue one’s vision are critical—and in China, he found opportunities that aligned with his long-term research goals.
