Seeking to outsmart US, China races ahead on artificial intelligence
Seeking to outsmart US, China races ahead on artificial intelligence
BY TIM JOHNSON February 21, 2018 05:00 AM Updated February 21, 2018 03:13 PM
WASHINGTON When a Google computer program beat the world’s best player of an ancient Chinese board game last May, it might have seemed like an incremental milestone.
But for some, the success of the program known as AlphaGo marked more than a man vs. machine clash. It set up a broader race between China and the United States over artificial intelligence, a competition that could mold the future of humankind just as the widespread arrival of electricity did in the last century.
The Go tournament took place in Wuzhen, a city of canals that is more than 1,300 years old, a fitting venue for a competition involving the strategy board game Go that has been played for several thousand years. Go is renowned for its complexity, and it is said that there are more variations to the game than there are atoms in the universe.
Perhaps it was a coincidence of timing, but the AlphaGo competition kicked off events that demonstrated China’s resolve to close the gap with — and quickly surpass — the United States in deploying artificial intelligence, or AI. Goals Chinese authorities announced last July are ambitious: Reach parity with the United States by 2020, achieve major breakthroughs by 2025, and “occupy the commanding heights of AI technology by 2030” as the world’s undisputed leader.
Can China do it? Experts say the race is in its early stages but the challenge has been set, and China is taking action to move toward its aspirations.
“There’s a lot of ambition, a lot of enthusiasm but it still remains to be seen whether this is possible,” said Elsa B. Kania, a specialist in artificial intelligence and Chinese defense innovation at the Center for a New American Security, a think tank in Washington. Nonetheless, she added, “there is a very real chance and possibility” that China could achieve its goals.
The stakes are high. Advances in artificial intelligence could add trillions of dollars to a major economy and give an edge on the battlefield, shifting empires and global power.
“For the moment, the United States is the most advanced AI country in the world. But that gap is closing,” said Chris Nicholson, chief executive of Skymind, a San Francisco start-up that focuses on deep learning, a type of artificial intelligence.
Russia, too, is paying attention, although it is not in the same tier as China and the U.S.
“Artificial intelligence is the future, not only for Russia, but for all humankind,” Russian leader Vladimir Putin told students Sept. 1, according to the state-run RT network. “Whoever becomes the leader in this sphere will become the ruler of the world.”
Processors equipped with artificial intelligence algorithms and specialized hardware platforms learn from experience, adapt to new information and perform human-like tasks. The field is in its relative infancy, and numbers of top scientists are limited. Google, the Mountain View, California, technology giant, employs at least half of the top 100 of them, the Eurasia Group, a New York-based political risk consultancy, estimated in a report in December.
But China has some comparative advantages, and it is moving fast. Last year, its scientists sought 641 patents related to artificial intelligence, compared with 130 in the United States, says CB Insights, a venture capital database company.
Other metrics also show China’s push. Every year, top experts gather for the Association for the Advancement of Artificial Intelligence. In 2012, U.S. researchers presented 41 percent of the papers at the meeting, and Chinese experts barely hit 10 percent, said Avi Goldfarb, a professor of marketing at the University of Toronto’s Rotman School of Management.
In 2017, he said, U.S. share of papers fell to 34 percent and China’s grew to 23 percent.
“What’s really happened is that China has invested a lot and done much better,” said Goldfarb, co-author of Prediction Machines: The simple economics of Artificial Intelligence.
One of China’s comparative advantages in artificial intelligence is the massive data that the government and large internet companies collect on the nation’s 1.3 billion citizens.
In China, people use their mobile phones to pay for goods 50 times more often than in the United States, says Sinovation Ventures, an early stage venture capital firm with offices in China and the United States. Chinese order 10 times more food delivery volume than in the United States. And shared bicycle usage is 300 times that of the United States, the company says.
“All of that is data that goes back to train AI models,” Kai-fu Lee, Sinovation’s founder and a former Google executive, told an MIT forum Nov. 2 on the future of work.
A Chinese bike-sharing company, Mobike, has fleets in 200 cities globally. Its sensor-equipped bikes transmit 20 terabytes of data for Mobike to analyze with artificial intelligence every day.
China’s government has moved heavily into facial recognition software designed to keep tabs on the populace. It has done so through unique levels of partnership with private Chinese tech companies, like Yitu Tech, one of many big firms that cooperate closely with authorities.
Lee, the former Google executive, said massive datasets give China an upper hand.
“A very good scientist with a ton of data will beat a super scientist with a small amount of data any day,” Lee told the MIT forum.
Given such an advantage, venture capital money is pouring into China-based startups pursuing aspects of artificial intelligence. Such startups took 48 percent of all dollars going to AI startups globally in 2017, surpassing the United States for share of dollars, CB Insights says.
While competition to harness artificial intelligence is real, it is unlike the race to develop a nuclear bomb before and during World War II. A remarkable degree of commercial overlap occurs between the two countries, and some experts don’t see an inevitable winner and loser in the competition.
“There’s a lot connectivity between Beijing and Shanghai and Hangzhou and (Silicon) Valley,” said Paul S. Triolo, director of global technology at Eurasia Group.
Microsoft and Google have large operations involved in artificial intelligence research in China, he said, “and Chinese companies like Baidu and Alibaba and Tencent have lots of AI researchers working in the U.S. Who’s winning that?”
“It’s not a race to build the first nuclear weapon. It’s a set of very diffuse capabilities that are used as part of bigger systems,” Triolo said.
The extent of how artificial intelligence might impact everyday life is not yet known. One prominent expert, Andrew Ng, a former chief scientist at Baidu who now heads a research group at Stanford University, compares the transformative power of AI to electricity, noting how electrical refrigeration transformed agriculture, the electric motor remade manufacturing and the telegraph revolutionized communications.
“If it is like electricity, it can transform economies, societies and militaries in ways that we can imagine and ways that we probably can’t imagine yet,” Kania said.
The potential has seized the attention of the People’s Liberation Army, which makes little secret that it sees artificial intelligence as augmenting China’s power. Military commanders paid close attention last year to the contest surrounding AlphaGo, which was developed by a Google subsidiary, DeepMind, based in London.
“The PLA wants an AlphaGo for warfare,” Kania said.
One way that artificial intelligence may be deployed is in swarm warfare — the use of scores maybe hundreds of flying or undersea drones to disrupt enemy ships and submarines.
Such fleets of drones, harnessed with the right algorithms, could be used “to overwhelm and saturate very high-value weapons platforms,” Kania said, like aircraft carriers.
“We’re entering a new age of warfare where AI and robotics will be much more integral to military power,” she said, noting that battlefield action might become so rapid that “humans can’t keep up.”
China has put some of its prowess on display. In an exhibition Dec. 7, a swarm of 1,180 drones took part in a nighttime flyover light show over the southern city of Guangzhou, spelling out “hello” and “innovation” to visiting dignitaries. Intel, the U.S. tech giant, flew more than 1,200 illuminated drones this month for the opening of the Pyeongchang Olympic Games.
As China makes headway, some U.S. experts voice concern, even alarm.
“I’m assuming that our lead will continue over the next five years, and that China will catch up extremely quickly,” Eric Schmidt, then-executive chairman of Google’s parent, Alphabet Inc., said at a forum in November.
But Schmidt lamented what he saw as a lack of a U.S. governmental focus on how to forge ahead robustly in artificial intelligence.
“We don’t have a national strategy,” Schmidt said. “If you believe that this is important – as I suspect all of us do, but certainly I believe – then we need to get our act together as a country.”
Forecasters can only estimate the economic impact of artificial intelligence, but when they do the numbers are extraordinary. PwC, a global management consultancy, said in a study last year that artificial intelligence could add $7 trillion to the Chinese economy and $3.7 trillion to the North American economy by 2030.
It would impact areas such as self-driving cars, improved imaging diagnostics, better fraud detection, efficient energy consumption, greater tailoring of apparel and consumer items, and personalized delivery of entertainment.
Industry experts are appealing to lawmakers to sustain the current U.S. edge.
“AI is the biggest economic and technological revolution to take place in our lifetime. … We can’t afford to allow other countries to overtake us,” Ian Buck, vice president for accelerated computing at Nvidia, a Santa Clara, California, manufacturer of circuits used in super-fast processing, said in prepared remarks to a House subcommittee Feb. 14.
Nicholson, the San Francisco entrepreneur, said the Trump administration has cut research and development programs, blocked mathematicians from the Middle East and curtailed foreign investment in areas like artificial intelligence.
“The United States is adopting policies that will make it lose this race,” he said.