Superpower Stalemate: AI and the US-China Duopoly

AI cover.png

The evolution of artificial intelligence (AI) first began in 1956 at Dartmouth College, when the term was originally coined by computer scientist John McCarthy at the very first AI conference. It was there that the visionary McCarthy set out to create machines that could “use language, form abstractions and concepts, solve kinds of problems now reserved for humans, and improve themselves.” Through this first step, McCarthy gave birth to the field of AI and set the stage for the disruptive technology emerging today, and the science fiction future many expect it to create tomorrow.

More than 60 years later, AI is now one of the most impactful technologies emerging, pervading industry and daily life. Through its fervent innovation, funding, and development, artificial intelligence stands to disrupt all sectors and all manner of business models. The field has come a long way since 1956, with advanced capabilities and new applications that have already fundamentally changed the world. Through the explosion of data, advancements in AI algorithms, and the increase in computing power and storage of recent years, the next stage in AI’s evolution has truly begun. But, in comparison to what we will see in the next decade or two, we are still in the stone age of artificial intelligence.

Looking at both the historical development of AI, and at the global ecosystem within which it grows, there are many who stand beside John McCarthy in the pantheon of artificial intelligence visionaries, including entrepreneurs and innovators like Ray Kurzweil, Andrew Ng, and Jeff Dean. But the focus has switched to what advancements in AI will mean for us all. One man adamantly answering this question is the much-heralded Kai-fu Lee.

Lee, a noted innovator in the field of AI since the 1980’s, is most well-known for his time as Principal Research Scientist at Apple, his pivotal role in establishing Microsoft Research in China, and as the former President of Google China. Now, Lee is using his vision for technology and artificial intelligence to drive the field forward at Sinovation Ventures, where he is the CEO, and President of the venture capital firm’s Artificial Intelligence Institute. Lee’s understanding of artificial intelligence is obviously an asset to his investments, but more than that, it gives him the ability to assess the greater landscape – understanding that AI does not develop in a vacuum, and the profundity of its implications.

Tied For First

To this end, the venture capitalist has outlined, in no uncertain terms, what the advancing of artificial intelligence means for society, business, and the geopolitical environment. Lee’s primary focus lies with the supposed AI “arms race” between the US and China. Often referencing a forthcoming US-China duopoly on the technology, Lee has detailed a world where AI dominance belongs to only the two nations. And, through his expert explanations, its hard to disagree with his point of view. “Today, the US clearly leads in advanced technology and research and has the best universities, and China leads in having the largest amount of data from the largest market,” Lee explains.

Lee references that both nations have very strong examples of large tech companies, such as Google, Facebook; Tencent, Alibaba. And though there might be some other opportunities for other countries, the fundamental nature of AI will cement the US and China as the global leaders. Lee supports this explaining that, “AI is a technology area that is really reliant on two things: 1) A huge amount of data. 2) A supply of experts and engineers. In these two respects, the US and China dominate probably 80 percent of the talent, and at least more than the majority of the data.”

Other countries, he says, can form their specialties, but the leaders are essentially a foregone conclusion. “Israel is always very strong in core tech. Canada has a bunch of AI talent. France is declaring the world’s largest incubator for technology. Singapore welcomes global talent. India, Indonesia, Brazil also have large markets. There are ways in which they can play, but the true big leaders are more or less set in stone,” proclaims Lee.

Data and Culture

One of the distinct advantages that China has in terms of advancing artificial intelligence is the homogenous nature of its massive amounts of data. “The data comes from one population, that speaks one language, from one culture,” Lee points out. “For example, Chinese people who do ridesharing with Didi basically view the Didi platform and the way in which they use cars to get home in a relatively similar way. The language, the culture, the method of payment, the preference of cars versus bikes, walking, or public transportation – these are all well understood elements.

Didi has four times the usage of Uber in the US, and, as such, Didi collects significantly more data overall. While Uber does have a lot of data worldwide, China still has more data and the data collected in China is homogenous. This means the data can be easily integrated and is actually beneficial to the development of artificial intelligence. Contrastingly, data collected from different countries can’t so easily be merged together, nor can artificial intelligence use it to make predictions as if the routes, languages, payment methods, and local rules and regulations were the same.

“For Uber to operate in 100 countries, they have to build 100 companies that each establish their own practices and data. And, of course, Didi would have to do that too if they wanted to go to 100 countries, but one China is as large as 100 countries,” Lee explains. With such a massive population, the number of mobile device users, and the uniform methods of digital payment, such as WeChat Wallet and Alipay, China has turned into one of the most hospitable and advantageous incubators for artificial intelligence to flourish.

Data’s role in China’s AI advantage, however, is only one factor. Culture, too, plays a part in how AI is developing. “Take, for example, video cameras in shopping malls. I think more Americans would be concerned about that,” Lee points out. Building autonomous stores by capturing data in a shopping mall may prove very difficult in the US where a certain level of privacy is expected. Lee describes how Chinese consumers would be more amenable to such a circumstance as long as they can see the added-value. “Chinese people are more willing to trade off privacy for convenience. Once you are willing to let a camera follow you in a shopping mall, that will lead to less shoplifting, better targeting, better sales predictions. The wrong kind of stores will close, and the right kind of stores will open. What you want to buy will be clustered and there will be targeted sales,” explains Lee.

The China Advantage

Looking at both sides, Lee sees advantages in AI for both the US and China. “I think China’s benefits are data, and the large number of young engineers who want to rush into this industry. That’s the second advantage – the quantity of young engineers and scientists.” The third advantage, according to Lee, is the Chinese venture capital environment. Lee points out that Chinese VCs are pouring more money into AI than American VCs. Meanwhile, the VC environment in China differs to that of the US: In China, there are more people funding AI for a greater variety of applications, and this results in more experiments and iterations and creates more opportunities for success.

Surprisingly, Lee ranks the Chinese regulatory landscape as one of the least impactful advantages to China’s AI ecosystem. “I think many people like to deemphasize this, but I have to acknowledge that this is an element. I think many American readers would view this as the Chinese government playing a giant role in making the industry.” It’s really more commercial than governmental, Lee stresses, such as building a new city – Shouhan – to try out autonomous buildings or building a new highway in Zhejiang to make electrical and autonomous vehicles more market-ready.

The other aspect in which the government aids the development of AI is the approach it takes to regulation, which is more tech-friendly and techno-utilitarian. “Without finishing the debate on how to regulate a technology, the Chinese government tends to allow technologies to go out and develop, and then figures out how to regulate it later,” says Lee, something that’s been seen in the fields of e-commerce, payments, transactions, and electronic financing, among others.

This style is in complete opposition to what’s seen in the US. Unlike in China, the US government’s approach to innovation is very much regulate first, ask questions later. The American government puts “a lot of focus on, for example, slowing down the advancing of autonomous trucks to protect the truckers’ unions,” Lee describes. This issue of job displacement, however, is not one that the government can control for very long. The impact of disruptive technologies like AI is only going to speed up as the years go on, so preparing for this eventuality will be a greater service to more people, rather than just pushing it down the line.

The US Advantage

However, the US does have plenty of its own advantages. Lee explains that, “The American advantage, first and foremost, is the more innovative and creative, out-of-the-box ability to think of new things that haven’t been thought of before.” This is something unique to US innovators and entrepreneurs that can’t be found so easily in the rest of the world or at such scale. “If something merely requires building a product, fitting the market, tweaking to the market, I think the Chinese entrepreneurs have gotten better at it than their American counterparts,” says Lee.

Out-of-the-box innovation the likes of Google, Tesla, or Apple exemplify the American advantage, and demonstrate the impact such innovative capabilities have on the world –  driving the creation of things not yet thought of before, satisfying widespread markets not yet understood. The innovative culture capable of producing such forward-thinking companies and products is something distinct to Silicon Valley and will remain a formidable force for the foreseeable future.

Moreover, the advanced research institutions of the US are a distinctly American advantage which can’t so easily be supplanted. “In top universities, top researchers, top papers, Americans still dominate,” Lee says. “Additionally, the US funding agencies – NSF and DARPA – really lead the way if you look at the top applications for AI today. The most well-known being speech recognition, machine translation, macro language dialogue, and autonomous vehicles, all of which were DARPA moonshot projects 30-40 years ago,” he says.

Yet another American advantage in AI is the US education system. Lee has seen that the American education system “attracts the best people, smartest engineers, scientist, scientist-wannabees, engineer wannabees – flying across the world because they want to study in the US from undergrad to graduate, study under the best research guidance, in the best engineering programs, getting the best training.” The benefits of this international attraction are compounded by the fact that once the world’s best and brightest graduate from US universities, a large percentage of them remain in the US, working on high-level research and developing solutions to advanced problems. “While the US population of indigenous American talent has to be a bit smaller than Chinese because of the sheer population difference, America draws from the world’s best and that is the US’s advantage over the rest of the world,” says Lee. “The fundamental magnetic attraction for talent.”

Who Will Win the AI Arms Race?

Today, wild speculation surrounds the question of who will win an eventual AI arms race. Many believe that with China’s open attitude towards tech development and loose regulation, incomparable data generation, and willingness for data collection, that China has a clear advantage. Lee, however, doesn’t see it being so cut and dry.

“I think it’s hard to predict the outcome. The American free spirit of expressing individualism is the essence of the American innovative advantage. And, the government policy is consistent with permitting that kind of discussion in the space with respect to regulation, policy, and trying to reach a consensus before becoming law, and then going through due process. That’s kind of a fundamental part of what made and makes American great,” puts Lee

Both sides, however, are well matched in different arenas. “The Chinese side is about pragmatism and execution, getting results and fast decision making, and then making something successful,” he says. The regulation exists, the governing bodies are there, but pragmatism is key in allowing technology to develop. “It’s not like China blindly permits everything. It’s when the disadvantages outweigh the advantages that they are quick to change, and limit the damage. And that’s the Chinese advantage,” Lee says.

Lee says the purpose of his upcoming book, AI Superpowers: China, Silicon Valley, and the New World Order, is not to speculate over who will win any sort of AI arms race. Rather, it’s about educating both sides to have a comprehensive understanding about what might be on the horizon. “The reason the book is in English is that the information is a little bit asymmetric. Chinese people understand the difference between the systems a little bit better, because many Chinese have been to America. I think many American readers potentially haven’t realized how far China has advanced and assumed the Chinese system could not possibly scale or deliver similar results to America. The book is intended to show that two systems exist, and you may prefer one or seem more aligned with the American one, but I think people should keep their eyes open,” says Lee. “In this duopolistic future, both systems will be tested, and we as inhabitants of the earth should be open-minded and understand what’s happening.”

AI in Industry

“I think AI and big data have already had a profound effect on the whole internet industry. Next will be financial. That’s already begun to happen. That includes banks, loans, and insurance because there’s a large amount of data in those areas and there’s an immediate economic benefit that can be used to guide the AI and measure the efficacy of the AI,” says Lee.

“Another big one will be machine learning and speech interaction, because those are just on the cusp of being ready. That will start to remove the language barrier amongst people. And that will be a big deal. First from print and then from spoken language. It will probably take five to eight years to play out,” says Lee. Another area of disruption will be the use of cameras and sensors indoors to upgrade the quality of healthcare, retail, and education. Lee describes that such “use of cameras and sensors can lead to things like autonomous stores, helping bed ridden patrons not to fall, improving customized personalized learning, and starting to become tools for doctors to consult for diagnoses and comparisons with other patients’ diagnoses. I see these as the next wave.”

The Future of Work

Despite the potential for a duopoly on artificial intelligence between the US and China, and the surface level competition, AI brings up a lot of issues that the two countries should continue to monitor and learn about from each other. “In the book, the area I go into the most is AI job displacement. Because AI will increase in capabilities, there are strong reasons why it will cause challenges in both China and the US,” Lee explains. “In China, people talk about it a bit more, because AI will replace the routine jobs and China is the manufacturer of the world – the assembly line workers are going to be replaced.”

Similar concerns, however, can be found in the US. Though not on the same scale, job displacement in the US will be an area in need of addressing in the near future. “We should recognize that there are two issues in the West that will cause a pretty fast displacement,” says Lee. “One is that white collar jobs are immediately more replaceable than blue collar jobs. Think about a loan officer. That job is done completely by a machine without any mechanical parts – it’s just software. Whereas an assembly line worker, you must build a robot arm that functions, robot eyes which see. This still offers technical challenges,” Lee claims. “White collar jobs are potentially more prone to AI replacement sooner. But, eventually they’ll both be replaced. I’m not trying to debate which country is more prone, I think they’re both prone,” states Lee. “We face the same issues, and potentially the same fate if we don’t figure out how to respond.”

Some jobs will be more vulnerable to AI displacement in the near term than others. “I categorize jobs differently and talk about areas in which AI is weak, and those are areas where we should be training, and our education should be modified to prepare for this displacement,” says Lee.

To respond to these areas and the potential for displacement, providing the gateways and career paths for people to do the right things is imperative. Lee puts that, “We should try to find the creative people and give them the fastest path to realize their dreams. And then, follow their hearts and do what they love in terms of creating, whether it’s a novel or a script or a scientific breakthrough or new medicine. And then we should let people continue to do the strategically complex jobs, like that of CEOs, but enhance them with AI as tools, so they can make more right decision and generate more economic benefit.”

Creating the jobs to compensate for AI job displacement requires a better understanding of work. “There’s as predicable order in which human labor will be replaced by machine, so we should retrain people to the jobs that are least likely to be replaced.

As well, there are jobs that we as humans will simply never want replaced by AI. “People don’t want a robot nurse, tour guide, social worker, psychiatrist, or teacher. So, all of those will expand to raise the levels of compassion, EQ, and empathy required. So, we should train for more of those people,” says Lee.

“Creating the AI tools, bringing the awareness of AI to professionals, increasing the number of compassion jobs, changing education to cater to each of these areas – these are the kind of things that China, the US, and the rest of the world should do. And because the US and China will reap the greatest economic benefit from AI, they have some responsibility to experiment both philosophically as well as entrepreneurially to increase these jobs and to show the rest of the world,” advises Lee.

Lee believes that the US and China are pioneers from not only a technological aspect, but also from social and governmental perspectives. As such, he puts that there is a responsibility on both countries to lead the way as AI’s disruption begins to permeate all nations, and all walks of life.

The future, Lee believes, may be very different than most imagine because of AI’s monumental impact on society. “In 50 to 100 years people may well work very differently. Some people may just stay home and write poetry – for which they may or may not get paid. But that utopian future may be possible if – and only if – we get through the transition, because joblessness, or loss of meaning through joblessness, can create so much social chaos that we may never get to that utopian future.”


This article was originally published in the pilot of the upcoming AmCham China Quarterly, an executive-targeted periodical focused on policy, business, and technology, driven by C-suite perspectives and insights. The AmCham China Quarterly is set to launch early 2019  to subscribe or contribute to the Quarterly, contact our editor:

For more news, events, information, and articles from the T&I Initiative, follow our LinkedIn showcase page.