The following article summarizes an interview of Chinese AI expert Song-Chun Zhu, who argues that China’s AI industry should chart a different course than the current U.S. focus on data- and compute-heavy large language models. He contends that China should explore multiple paths to general purpose AI simultaneously, such as modeling human cognition, algorithm innovation, and “small data.”
An archived version of the Chinese source text is available online at: https://perma.cc/4KUL-NGGF
Song-Chun Zhu: The Race to General Purpose Artificial Intelligence is not Merely About Technological Competition; Even More So, it is a Struggle to Control the Narrative
On January 11, the 26th Peking University Guanghua New Year Forum was held at the Centennial Lecture Hall of Peking University. During the forum, Song-Chun Zhu (Zhu Songchun; 朱松纯), a member of the 14th National Committee of the Chinese People’s Political Consultative Conference (CPPCC) and dean of the Institute for Artificial Intelligence and the School of Intelligence Science at Peking University, delivered a keynote speech and stated that independence of thought (思想自主) is the prerequisite for science and technology (S&T) innovation, while confidence is the core driving force for promoting high-quality economic development. In an exclusive interview with Economic View, Zhu specifically emphasized the urgent need to reform existing evaluation systems and incentive mechanisms, implement the original strategy of “playing Go” (“下围棋”), and focus on the overall situation rather than isolated victories.
Independence of Thought in S&T Shapes the Narrative System
Confidence is an “invisible variable” in current economic and technological development. Zhu highlighted that building self-confidence requires robust technological support and, more importantly, building one’s own narrative system, which requires “independence of thought” in S&T innovation.
“The development of general purpose artificial intelligence (“general AI”)1 depends on big data, computing power (compute), and chip technology. In the background, however, is not merely about competition in technological capabilities but, even more so, it is a struggle for global confidence and for control of the narrative” (叙事权). Zhu pointed out that in the field of S&T, there has always been a “U.S.-dominated narrative.” For example, in AI, the promotion of this narrative has solidified the United States’ dominance in general AI by emphasizing that big data, compute, and chips are insurmountable barriers for other countries. At the same time, the market capitalization of U.S. technology giants in the U.S. stock market has soared.
This type of sentiment has far-reaching implications, creating a “confidence gap.” On the one hand, the confidence of individuals and enterprises in other countries has been undermined, leading to conservative investment decisions and slower innovation. On the other hand, international capital flows to the United States, continuously funding its technological research and industrial upgrades, thereby widening the technological gap. In such an environment, the United States has consistently gained the upper hand in global technological competition.
To view the rest of this translation, download the pdf below.
Download Full Translation
Song-Chun Zhu: The Race to General Purpose Artificial Intelligence is not Merely About Technological Competition; Even More So, it is a Struggle to Control the Narrative- Translator’s note: The Chinese term 通用人工智能 can be translated into English either as “artificial general intelligence” (AGI) or “general purpose artificial intelligence” (“general purpose AI,” for short). This translation opts for “general purpose AI” because when Chinese writers use the term 通用人工智能, they are usually referring to a broadly capable form of AI, rather than to AI that resembles human cognition, as AGI implies. For a more thorough discussion of this term, see Wm. C. Hannas, Huey-Meei Chang, Daniel H. Chou, and Brian Fleeger, “China’s Advanced AI Research: Monitoring China’s Paths to ‘General’ Artificial Intelligence,” Center for Security and Emerging Technology, July 7, 2022, https://cset.georgetown.edu/publication/chinas-advanced-ai-research/, pp. 1-3.