Follow Datanami:
August 19, 2019

U.S., China Jockey for AI ‘Lead’

via Shutterstock

The U.S. is maintaining its lead in AI development as China hustles and spends to catch up, outpacing its rivals in areas like AI funding, adoption and access to data.

Still, a report released this week by the Center for Data Innovation finds the U.S. remains the clear leader in key categories such as R&D, “talent” and hardware development. China still trails the U.S. despite heavy investment, with the European Union trailing, paying what the report said is an “economic price [by] staying on the sidelines” of the AI race.

The current American lead in AI development stems largely from a vibrant AI startup community that has attracted heavy venture capital and private equity investment.

Another key contributor is the huge U.S. lead in AI chip development by Nvidia (NASDAQ: NVDA) and other chip startups. For example, Cerebras Systems unveiled a dinner plate-sized “wafer scale engine” on Monday (Aug. 19) aimed at accelerating AI processing.

While the U.S. publishes fewer research papers than China and the EU, those results are deemed of higher quality, as is the overall talent among U.S. developers, which the report classified as “elite.”

On its 100-point ranking methodology, the industry group pegged U.S. AI efforts at 44.2 points. China followed with 32.3, with the European Union trailing with 23.5.

China is nevertheless closing the gap with the U.S., with greater access to data—much of it gathered through surveillance—used to train machine learning models. The rivals split the difference when it comes to AI investment, with the U.S. regaining the lead in venture funding last year, according to Washington-based group’s report.

Source: Center for Data Innovation

Where China lags is talent among its AI developers. For example, the report notes that Italy alone had more AI researchers ranked in the top 10 percent internationally than China in 2017. Among the remedies are expanded training and educational efforts and fostering a “stronger culture of open data,” the report concludes.

Indeed, researchers have noted that China’s AI ambitions may be stunted by strict government controls on data access. It remains difficult for developers and researchers to “get access to this big wealth of data sitting in government agencies,” Xiaomeng Lu, international public policy manager for technology consultant Access Partnership, noted during a panel discussion last year hosted by the Center for Data Innovation.

China’s ambitious AI strategy has it leading the world by 2030, a goal Lu characterized as “aspirational.” Chinese AI researchers face not only lack of access to domestic data but also government “firewalls” that prevent them from tapping into international data sets, added Lu. “Only the free flow of data can generate value.”

The AI report also emphasizes the economic and national security imperatives related to AI development. “Nations that lag in AI adoption will see diminished global market share in a host of industries, from finance to manufacturing to mining. And nations that underinvest in AI R&D, particularly for military applications, will put their national security at risk.”

“The United States is leading in AI today, but it should not rest on its laurels,” said Michael McLaughlin, the Center’s research analyst and the report’s lead author. “To maintain its lead, the United States should focus on policies that grow its domestic talent base, enable foreign AI talent to immigrate and increase incentives for research and development.”

Skeptics question the utility of handicapping an AI race. “Clearly, many nations have experienced an explosion in AI funding, startups, patents, research, hiring, degrees, product and so on,” James Kobielus, a lead analyst with SiliconANGLE, wrote in Datanami earlier this year. “But none of these metrics can be used to definitively declare anybody’s ‘leadership’ in this multifaceted arena.”

Recent items:

China Ahead on AI Strategy, Behind on Data Access

Let’s Accept That AI Leadership Is Everywhere

 

 

Datanami