Algorithms trained on massive amounts of Chinese data may soon make decisions that profoundly affect American lives.
Take Yitu Technology, a Shanghai-based AI startup that won top honors in two AI competitions in the US last year, thanks to its advanced facial recognition technology. It is understood that the system was built for Chinese law enforcement using data collected by Conquest. Or as the company calls itself, "the world's largest portrait system, covering over 1.5 billion people."
Currently, Yitu is now looking for customers in the US to get its software to work. "There are many applications for this technology," said Wu Shuang, who heads the Silicon Valley research group . The company is just one simple example of the growing competition between the United States and China in the field of artificial intelligence, and of course it's not the only one.
Another Shenzhen-based Malong Technologies has also used massive amounts of Chinese data to train image recognition algorithms, such as analyzing thousands of photos from fashion shows, to determine trends in customer demand in the apparel industry. The company said it is now working with U.S. e-commerce companies to test the technology.
Matt Scott, the company's founder who was also a former researcher at Microsoft, said: "In China, the main difference is that there are more people, more data, more companies -- and to sum it up, the scale has become Even bigger, and after China gets this data, we can export (the technology) around the world.”
In the age of big data, there is still a long way to go to determine who is the real leader, but algorithms like this can be described as the vanguard of the competition, and China is quickly catching up and becoming the United States in this competition strong opponent.
The AI revolution is often thought of as robots or drones because they can perform some of the tasks that were once done by humans. But its effects will also be felt from a less obvious source: data analytics. Models of machine learning systems can be found by analyzing large datasets, a technology that is now at the forefront of artificial intelligence.
For some industries, deep learning -- the most advanced form of technology -- has the potential to create value equivalent to 9 percent of a company's revenue, in other words, the equivalent of tens of thousands of dollars, according to an April McKinsey Global Institute report. billion in potential economic value, of which the United States and China are the clear leaders.
"If you look at the world, it's a two-horse race in artificial intelligence," said Michael Chui, a McKinsey partner who led the research.
In China, the artificial intelligence boom has boosted the country's self-confidence on the basis of its ever-improving technology. In China's latest plan, artificial intelligence has become one of the central pillars of the "Made in China 2025 plan," which aims to transform the country's economy and set a goal of becoming a world leader in technology by 2030.
At the same time, China's progress has created another paranoia in the United States, that its technological exceptionalism is no longer taken for granted. So now the Trump administration's motivation to plan a trade war with Beijing is at least partly driven by concerns about China's advances in new technologies.
“It’s clear that the U.S. government sees itself in a technology arms race with the Chinese government,” said Robert Silvers, a partner at law firm Paul Hastings and former assistant secretary for cyber policy at the Department of Homeland Security. "The United States sees these technologies as transformative, so that leading countries have not only an economic or technological advantage, but also a national security advantage."
One reason AI is so competitive is that it is tied to a race to find new military advantages. In addition to answering questions from ordinary users and applications in the field of unmanned vehicles, the technology can also synchronize swarms of drones, analyze photos taken by spy drones, and control autonomous boats.
Sean Gourley, founder of Silicon Valley AI startup Primer, said the advantages of AI could bring about a change in warfare. The company's backers include the CIA's venture capital arm. One thing we have to admit is that technological shifts like this could erode the military advantage of great powers.
“This could affect the reordering of global power. Whoever does it better is going to be in a weak position for almost 10 years,” Gourley said.
Last year, Russian President Vladimir Putin inflated the bet on AI in his own way: "Whoever becomes the leader in this field will become the ruler of the world."
However, according to most experts, the United States still has a clear lead. To accomplish world-class artificial intelligence, three things are needed: the most advanced algorithms, the most specialized computing hardware, and machine learning raw materials.
In last year's Go game, AlphaGo, created by Google's DeepMind, defeated Chinese leader Ke Jie. A Google executive has said this is a wake-up call for China in the field of artificial intelligence. "It was only when the Russians launched Sputnik that the U.S. realized how far they had come; similarly, it was a similar experience when the Chinese Go champion lost to AlphoGo."
This view is shared by some in China. “The U.S. is obviously still the main resource for top talent. I don’t think there’s any doubt about that,” said Rong Jin, director of machine intelligence technology at the Dharma Institute, an Alibaba research program. In China, some believe that Americans are actively engaged in basic research, and that many academics are heavyweight mathematicians—a core subject of artificial intelligence—while the Chinese tend to study coding or engineering.
But despite these advantages, China is now rapidly closing the algorithm gap. When it comes to the output of Chinese research institutes, Oren Etzioni, director of Microsoft co-founder Paul Allen's Artificial Intelligence Institute, said, "The statistics are definitely rising sharply." He also pointed to other signs of China's AI growth , For example, in a reading test earlier this year, artificial intelligence "newcomers" Alibaba and traditional forces sent Microsoft both won the highest honors. In addition, Chinese researchers also performed well in the annual image recognition competition.
On the other hand, in the field of software development, China has made slow progress in building a local chip industry, which needs to be placed in a major research position. That's partly due to a series of decisions that effectively blocked their takeovers of U.S. chip companies. This decision, which began with the Obama conquest, is now being accelerated by the Trump administration.
As for the last aspect, the availability of raw data. Most experts believe that China's AI advantage lies in this.
China has a lot of data on its citizens and is good at using it. Facial recognition technology is so widespread in China, for example, that it can be caught running a red light while crossing the road, or even observed stealing in public.
But it's also a continuation of the early development of China's internet: When people shop online, pay for money, and play games, they leave behind data that allows merchants to survive. A Chinese machine intelligence scientist once said: "The density of people is proportional to the density of data."
China is a little more relaxed about data privacy, but in Europe it's at the other extreme. The region will introduce strict privacy rules, the General Data Protection Regulation ( GDPR), later this month. However, U.S. companies like Facebook, Google and Amazon also hold vast amounts of data, said Wu Shuang of Yitu Technology.
This suggests that general AI applications such as facial recognition will be the preserve of all "big platforms" regardless of their country of origin, said McKinsey partner James Manyika. In contrast, if the data is rich enough, more specialized applications can be refined. In the manufacturing sector, for example, China is "gathering more data."
According to some experts, these data advantages may be greatest in areas where regulation makes it more difficult to obtain information or prevents information collection. Earlier this year, Google released a study showing that it could predict heart attack risk by using image recognition software to study retinal blood vessels. The study relied heavily on the UK's Biobank, a database that has been conducting detailed studies of UK volunteers since 2006.
However, only 631 people in the biobank had medical conditions associated with this study. This makes the dataset "relatively small for deep learning," reducing the effectiveness of algorithms trained with that information, Google said. In contrast, researchers in China's smart medical field can take advantage of larger datasets, one expert said.
If China has a wealth of data, it can also use its own economic opportunities to help attract more turtles. Artificial intelligence is being harnessed in the legal field, where machines have replaced more than 6,000 courtroom stenographers; managing traffic on roads; detecting tumors in hospitals; and in Shanghai subway stations, where passengers can buy tickets by talking to machines.
"Artificial intelligence has the biggest opportunity in China compared to other Western countries," Rong Jin said.
There are also plans for smart cities in China, which is different from the United States: large-scale cooperation between the state and private companies. In addition to projects aimed at traffic management, crowd control, finding missing children and the elderly, and reducing hospital wait times, all of the big tech companies have joint research labs with the government.
Wu Shuang said this is part of a lack of broader trials in the United States. “In general, China’s tech scene is more dynamic now, especially in terms of trying out new ideas and new products.” A leading Silicon Valley venture capitalist bluntly pointed out the difference in AI: “China’s The business is getting bigger.”
This economic impetus behind AI is closely aligned with the second most powerful force, a sense of national mission that brings not only funding but also clear industrial policy. This state-led strategy has also worked closely with private companies such as Baidu, Alibaba and Tencent.
And the US has done less to advance the national agenda. "Where is the national direction for AI and robotics in the U.S.? There's nothing. That part is missing, and the government is hovering."
To make matters worse, the Trump administration’s attempts to limit immigration have rattled the U.S. tech industry, which relies heavily on overseas talent — especially engineers from India and China. The AI chiefs at Apple, Facebook, Microsoft, and Google's cloud computing division are all from overseas.
"We're seeing more and more students choosing not to come to the U.S.," Etzioni said. “We are digging our own graves.” He pointed out that the pendulum of talent is moving away from the United States: Google and Microsoft have opened artificial intelligence research centers in China to tap artificial intelligence talents.
However, America's professional advantage will not disappear overnight. Companies like Yitu Technology are moving in the opposite direction, as they believe the US West Coast still attracts many of the world's top engineers. "Half of the AI engineers in Silicon Valley are Chinese," Wu Shuang said.