Login ID:
Password:
Partner Login
Contact Us : 7066511911

U.S. and Chinese Companies Race to Dominate AI -- WSJ

23 Jan 2018 7:32 am
By Sam Schechner, Douglas MacMillan and Liza Lin 

This article is being republished as part of our daily reproduction of WSJ.com articles that also appeared in the U.S. print edition of The Wall Street Journal (January 23, 2018).

Last fall, a U.S. intelligence-community contest sought algorithms that could identify surveillance images of people out of a database of millions of photos. The dark-horse victor was Shanghai-based artificial-intelligence startup Yitu Tech.

The five-year-old company beat out 15 rivals for a $25,000 prize. But Yitu had an advantage: access to Chinese security databases comprising millions of people that it could use to train and refine its algorithms.

Competitors, who worked from smaller databases, had "one hand tied behind our back," says Paul Nicholas, chief executive of CyberExtruder.com Inc., a Newark, N.J., startup that entered the contest.

The contest represents a growing concern about the global race to develop and profit from artificial intelligence, which frequently uses large data sets to learn skills like facial recognition and cancer diagnosis. Big U.S. technology companies are leading the race. But their Chinese rivals are catching up quickly because of growing investments, as well as freer access to enormous amounts of data about people, often compiled with the help of government agencies.

In the West, access to such information is at times limited by growing concerns over privacy and the ethics of letting machines make important decisions -- fears that are leading to new policies and stricter proposals around collecting personal data and deploying AI.

"Ultimately, AI advances are inextricably founded on the broad use of data to train machines as they go about learning," says Brad Smith, president and chief legal officer for Microsoft, which has invested heavily in AI and also is a vocal supporter of strong privacy protections. "The question is whether privacy laws will constrict AI development or use in some parts of the world."

An accelerating effort

At stake is whether U.S. giants -- including Apple Inc., Google parent Alphabet Inc., Amazon.com Inc., Facebook Inc. and Microsoft Corp. -- can retain their lead in the global tech race or whether they will be overtaken by Chinese behemoths Tencent Holdings Ltd. and Alibaba Group Holding Ltd. Also in the hunt is Chinese search engine Baidu, which is smaller but has invested heavily in AI.

When it comes to academic and corporate AI research, the U.S. -- particularly in the San Francisco Bay Area -- is a powerhouse. The U.S. dominates the number of international patents filed for AI topics like neural networks and unsupervised learning, according to filings from the World Intellectual Property Organization. The country also has more than 1,000 AI companies, almost double the number in China, according to a December report from Tencent's in-house research institute.

In addition, spending by U.S. companies dwarfs that of China's tech titans. Alphabet spent $13.9 billion on R&D in 2016, and Microsoft spent $13.0 billion for the year ended June 30. By contrast, Alibaba pledged in October to nearly triple its R&D spending -- to about $5 billion annually over the next three years.

But China hopes to close the gap with the U.S. in funding and tech development, driven by a new government effort to lead the AI field by 2030. The strategy focuses on boosting areas including medicine and agriculture, as well as government and military applications.

The drive is showing up already in the funding for AI startups. Among these Chinese firms, disclosed funding rose 10-fold in 2017 over the prior year to $7.06 billion, surpassing comparable U.S. firms, where disclosed financing rose 24% to $5.62 billion, according to CB Insights.

PricewaterhouseCoopers predicts that the U.S. will profit most from AI initially, but that China will catch up, reaping about 46% of the $15.7 trillion it expects AI to contribute to global economic output by 2030, followed by North America with 24%.

"The Chinese government's thoughtful investment in AI is a huge accelerator," says Andrew Ng, a former executive at Google and Baidu, who recently founded AI startup Landing.ai.

Different values

Another issue that looms large in the rivalry between China and the U.S. is privacy. Chinese consumers have fewer concerns about data privacy compared with those in the West -- and that gives Chinese companies greater leeway to mine their personal information and use it to develop smarter and more effective AI systems, for jobs such as medical care and surveillance.

For instance, Yitu, the AI contest winner, works on behalf of multiple security agencies in China and has access to a database of 1.8 billion photos of faces. Using that trove of information, it has learned to scan photos and find matches in seconds, to do jobs like authenticating the identity of ATM users.

Western policy makers are more reluctant to give companies wide latitude to use consumers' personal information, which some experts think could end up hampering the development of AI and let Chinese companies pull ahead in the tech-development race.

And privacy restrictions will get a lot tougher in May, as a new European Union law that covers tech titans like Google takes effect. The law limits the use of identifiable information about people, such as name, address and ethnicity, and allows individuals to demand explanations for automated decisions on important topics, such as why a computer denied someone's loan application -- answers that are difficult for some types of AI to provide.

More than 120 countries have so far adopted privacy laws, the large majority with European-style elements, according to Graham Greenleaf, a professor of law and information systems at Australia's University of New South Wales. Tech executives say Europe's size could force U.S. tech companies that wish to do business in Europe to follow its lead in other areas, too.

Regulation could be "a big help or a big detriment" to U.S. AI efforts, depending on how it's implemented, says Mr. Ng. "Despite the U.S.'s current lead in basic AI research, it can be easily squandered in just a few years if the U.S. makes poor decisions."

Mr. Schechner, Mr. MacMillan and Ms. Lin are reporters in The Wall Street Journal's Paris, San Francisco and Shanghai bureaus. Email sam.schechner@wsj.com, douglas.macmillan@wsj.com and liza.lin@wsj.com.
 

(END) Dow Jones Newswires

January 23, 2018 02:32 ET (07:32 GMT)

Copyright (c) 2018 Dow Jones & Company, Inc.
Disclaimer
Top 5 Special Reports
Weekly: Pulses Slips On Listless Demand In Dals, Liquid...
USD/INR (Dec. 19) Bearish Short-term Trend: Retestin...
USD/INR (Dec. 19) Retesting Major Support Near 71.00...
USD/INR (Dec. 19) Retesting Major Support Near 71.00...
USDA Revises Down 2019-20 Global Cotton Production Esti...