On October 14, 2022, CSCC had the pleasure to welcome Dr. Nan Jia, Associate Professor of Management and Organization at USC’s Marshall School of Business. Professor Jia presented her latest working research tackling the topic of AI Innovation in the United States and China, a topic that is highly relevant and anticipated, which was illustrated by the thoughtful engagement from professors and students alike in the Q&A section.
Professor Jia’s research touches on the broader question that is: what is the role of the state in AI development? And, does state involvement in innovation help, or take away from research progress? Professor Jia examines patent data from 1985 to 2018 between the United States and China, focusing on disparities between university-led innovation and firm-led innovation. Jia finds that while universities lead AI patents by a significant number in China, the equivalent is not true in the United States. In 2006, university-led AI patents produced in China diverged firm-led patents, coinciding with the Chinese national campaign for indigenous innovation, a government-led public campaign initiative to promote innovation among universities. When looking at share of university patents among provinces, inland provinces, where the commercial market for AI is small and firm activity is weak, had the highest shares. Among Chinese universities with high number of patents are world-class institutions such as Tsinghua University, but also on the list are lesser-known universities not typically renowned for cutting-edge technology innovation. On the other hand, US universities lead innovation in non-AI technologies, but are lackluster in the AI field. Jia suspects that while innovation in other sectors depend on resources such as lab technology, precision measurement processes, and other niche requirements, the critical enabler for AI innovation is data, which outweighs any other factor. In the United States, data lies in the hands of big technology companies such as Facebook, Amazon, Google, and Apple. Similarly, these companies are the ones making advancements in fields such as pattern recognition, machine learning, and neural networks, which are the leading categories of AI subsectors. In terms of industrial policy, the data factor raises a new question: what drives AI innovation? Is it funding, or is it access to data? Though most likely a combination of both, Jia hints that it is unlikely AI innovation only depends on data. Chinese universities, which have put out record number of patents, struggle to obtain comprehensive and workable data, which is tightly secured by the CCP.
Though CCP-led university innovation programs have loosened those restrictions somewhat, data transfer between government and university is slower than expected. Nevertheless, Jia does point out that AI industrial policy, whether the focus may be funding or data, is hard to get right. The challenge in crafting successful industrial policy is in allocating funding to the right places, and providing data in a secure way to universities without risking that data leaking out. Thus, it might be fair to say that state intervention in AI innovation may not net the improvements Western nations fear, and that, perhaps, the most successful method to foster innovation is a free-market approach. China is certainly working on finding the best approach, and it would be wise to observe their method. Professor Jia’s research provides much-needed guidance to this point, and we look forward to the completion of her research.