Advanced Computing in the Age of AI | Monday, January 24, 2022

Report: R&D Funding Will Help U.S. Retain AI Lead Over China 

via Shutterstock

Reports of China’s advantages in AI development and data access have been greatly exaggerated, concludes a report advocating greater U.S. investment in basic research and a reassessment of how to manage risks posed by Beijing’s technology push.

“China’s exalted advantage in AI is over-rated,” asserts the report released by the 21stCentury China Center at the University of California at San Diego. The U.S. and its western allies maintain advantages in key areas fundamental to AI innovation, including semiconductors and cloud computing.

“China is a formidable player in AI, but by most measures, the United States still leads, and draws on a set of strengths that China lacks and is unlikely to acquire in the near future,” concludes the report released on Monday (Nov. 16).

The one area where China does lead is facial recognition technology, which the report notes has been heavily funded by the government “for social control purposes.”

The study’s authors argue the U.S. maintains “asymmetric strengths” in AI development, including attracting talent from China. An estimated 80 percent of Chinese students earning degrees in the physical sciences remain in the U.S.

“We should more like us, we should double down on our asymmetrical strengths,” said Jason Matheny, a co-author and director of the Center for Security and Emerging Technology at Georgetown University.

China’s large population and lax data regulation have also promoted myths about its “data advantage” in AI development. That has led some observers to pronounce data the “new oil.”

Data is not “a general-purpose resource,” noted Matheny, who also serves on the National Security Commission on Artificial Intelligence. The federal panel has called for NATO allies to coordinate their AI efforts to ensure interoperability.

“While it is true that data are an important input to AI systems, one particular set of data is not generically useful for training any kind of system,” adds the report. “Any given AI application requires data relevant to the specific problem it is trying to solve.”

In order to sustain its lead. the report further calls for substantial U.S. funding increases for fundamental AI research as well as foundational semiconductor technologies. Chip investments should focus on next-generation IC design and manufacturing that would serve as the basis for future AI deployments.

The report recommends doubling the fiscal 2021 budget request to $50 million at the National Institute of Standards and Technology. Those funds should target AI testing, evaluation, verification and validation, the report states.

Overall, basic research for AI development should grow to $2 billion annually, the report recommends.

That total would reinforce another western advantage over China in the AI race: The U.S. and its western allies collectively invest twice as much annually on AI research.

“Heal thyself,” added Peter Cowley, dean of UC-San Diego’s School of Global Policy and Strategy. The goal of basic R&D for AI, 5G and other emerging technologies should be, “Get to where the puck is going, not where it has been,” Cowley said.

The report echoes earlier assessments of China’s technological prowess that characterized its national AI strategy as primarily “aspirational.”

About the author: George Leopold

George Leopold has written about science and technology for more than 30 years, focusing on electronics and aerospace technology. He previously served as executive editor of Electronic Engineering Times. Leopold is the author of "Calculated Risk: The Supersonic Life and Times of Gus Grissom" (Purdue University Press, 2016).

Add a Comment