China, America and the race for AI supremacy

It’s not just a tit-for-tat trade war between China and America that is currently escalating – a technological arms race in artificial intelligence is also moving rapidly and it’s largely a two-country competition.

A recent report by Oxford University titled ‘Deciphering China’s AI Dream’ found that, at present, China trails the US in every area of artificial intelligence development – hardware, research, algorithm, and industry commercialisation – except for access to big data. “I think the US is pretty clearly ahead now in terms of AI advancement,” says Jeffrey Ding, researcher at the University of Oxford. “As per my AI Potential Index, the US has 33% of the world's AI potential, whereas China trails with 17%.”

While the US might currently have the edge, the jury is out on who will win in the long-term. “It is too hard to predict, because the different drivers (hardware, data, research and development talent, and the commercial AI ecosystem) feeding into AI are rapidly changing and the relative importance of the drivers themselves is uncertain,” says Ding.

While the US most likely has the world’s leading researchers and talent in artificial intelligence, experts believe that China, armed with strong government support, three times as many internet users as the US has total population, over a billion smartphones and fewer data privacy laws could ultimately creep past its competitor in the race for AI advancement as soon as 2020, some predict. “The US is still by far the best in terms of top-tier talent, but China is closing the gap fast,” said Lu Qi, vice-chairman and chief operating officer of Baidu, ‘China’s Google’, at the CES technology trade show last year in Las Vegas.

In some cases, US companies have felt the need to withhold technologies from the Chinese market in order to protect their intellectual property

During a talk at a tech summit last November, former Google CEO Eric Schmidt told his audience that it wouldn’t be long before China overtakes the United States in the development of advanced artificial intelligence. “They are going to use this technology for both commercial and military objectives, with all sorts of implications,” said Schmidt, according to a report by The Verge.

China has made no secret of its ambitions to become the world leader in AI by 2030, investing US$150 billion to make the country an “innovation centre for AI”. Four of China’s biggest tech companies — Baidu, Alibaba Group, Tencent Holdings, and iFlyTek – have been appointed as “national champions” to lead the development of AI innovation in four different fields: autonomous driving, smart cities, computer vision for medical diagnosis and voice intelligence.

It’s shaping up to be a competitive race, but a tricky one to navigate – especially for multinationals with offices in both countries. Earlier this year Google opened an AI research centre in Beijing and, according to a 2018 campus recruitment list, US tech giants like Microsoft, Google and Amazon are hiring entry-level AI engineers in China at lower salaries compared to the US. But how do the multinationals avoid in-fighting when it comes to sharing technology and ideas?

“Companies that set up offices or research centers in both countries do so precisely to leverage the innovative capacity and access markets in those countries, which necessitates the sharing of technology and data across borders,” says William Carter, deputy director of the Center for Strategic and International Studies. “In some cases, US companies have felt the need to withhold technologies from the Chinese market in order to protect their intellectual property, but generally the exchange of technology and ideas at the working level is pretty robust.”

Self-driving cars in Beijing, in March 2018, after Baidu was given permission to start open-road tests

“It appears that in the tech domain most multinationals with offices in both countries have a pretty clear home base in one country over the other,” says Ding. “So Baidu, Alibaba, and Tencent may have overseas labs and institutes in the US, but it's pretty clear that the technology is based in China. Same goes for Google and Microsoft.”

Kirk Broody, an analyst with New Street Research, believes too much attention is given to the competition between China and the US when the real race for AI advancement is in-country. “We believe that in-country competition is more relevant. Alibaba is more concerned about Tencent than Amazon.”

Overall, Broody believes that issues around tech sharing are in the minority. “US companies are largely shut out of China whilst the China presence in the US is somewhat limited.” Meanwhile, “conflict on ideas is probably solved by people moving to different companies,” he says. 

Chris Nicholson, co-founder of artificial intelligence firm Skymind, believes multinationals with offices in both countries can avoid in-fighting by understanding very clearly the laws of the countries in which they operate, as well as business tendencies. “It is helpful to build alliances and partnerships to ensure that one has support wherever one operates.”

China takes [AI] seriously and the US does not. Not yet.

With a view from both sides, Nicholson understands more than most the overall level of competition between the two countries in advancing AI. He established Skymind in San Francisco in 2014, and has since expanded outside the US with an office housing four AI engineers in China’s Fujian and another office to follow next year in Shenzhen. Chinese tech giant Tencent is a big investor in his company. The competition between the US and China in AI is uneven, he says. “China takes it seriously and the US does not. Not yet. China experienced a Sputnik moment with computer program AlphaGo, while the US is counting on its private sector to produce cutting-edge research.”

Outside of the multinationals, the competition appears just as fierce with China currently taking the lead on investment in startups. According to popular technology analysts CB Insights, in 2017 China accounted for 48% of the world’s total AI startup funding, compared to 38% in the US.

However, while startups in both countries are staring to make some strides, it’s still the tech giants in both China and the US that, with access to big data, have an undisputed advantage.

“The big tech companies have a real advantage in terms of resources: not just hiring more R&D staff but also because of access to the very large data sets that improve output,” says Broody.

AI robots on display at the Appliance and Electronics World Expo 2018 in Shanghai

With the world’s largest online population and fewer privacy protection laws, China clearly has the advantage when it comes to harnessing big data. However, according to some in the field, this isn’t entirely good news. “In terms of application of AI technologies, China seems to have much less regulatory oversight than US does. While this is helping a faster proliferation of AI companies in China, it can also increase chances of misuse of AI technologies, with potentially adverse impacts on privacy and/or safety of the society,” says Subbarao Kambhampati, president of the Association for Advancement of AI.

However, while China’s tech companies like Tencent and Alibaba may have greater access to data than American companies such as Google and Facebook, experts like Ding believe that data alone is not going to be enough for the country to win the AI arms race. He says a lack of AI talent – experienced engineers and researchers – could hold China back. “Of all the factors, China’s struggle to train and attract the best talent could be the biggest setback for its AI dream.”

A report published last year by Chinese tech giant Tencent found that there are just 300,000 “AI researchers and practitioners” worldwide, but there is a market demand for millions of roles. Competition for the most highly skilled talent is the most intense. The report states that there are fewer than 1,000 people worldwide who are considered capable of steering the direction of AI research and development. The report also found that the US is currently “far ahead” in terms of global talent, with more universities doing AI research than any other nation, and more AI startups.

“The US has cultivated a culture of research where ideas and information circulate freely, and this accelerates the development of new technologies,” says Nicholson. “This has made the US a destination for talent from around the globe. That international talent has helped it become the world leader in AI. ”

The NextEV Nio EP9 electric supercar on display at an automobile exhibition in Beijing in December 2017. Nio, often called 'the Chinese version of Tesla' also launched its first in-car artificial intelligence system, called Nomi.

Either way, the ongoing competition will lead to “much more powerful AI,” believes Nicholson. “The country that makes the greatest commitment to the sector will advance more quickly, of course.”

It’s clear the battle to become the world’s leading AI superpower has become a two-horse race between China and America, with research, investment, and talent the main drivers. China has already made its ambitions clear: the country says it will become the world’s leader in AI by 2030. Stateside, former deputy defense secretary Bob Work and the former executive chairman of Alphabet, Eric Schmidt, have both urged the US government to respond to China’s national AI plan with a strategy of its own.

Meanwhile, experts like Ding remain optimistic that the race between China and the US to build better AI could benefit us all. “There are many ways for this competition to be healthy. One way for there to be a "race to the top" is in the domain of technical standards. If the international technical standards for AI require strict user privacy protections and robust risk assessment procedures for AI systems, then the US and China would compete to build the AI systems that match up with those standards. This could result in better predictions, better medical care, smarter objects and much more.”