The Race for Artificial Intelligence: China vs. America
What is The Role of Europe in the AI Race?
Let’s be clear, Artificial Intelligence, in particular in its latest development, deep learning that mimics the way the human mind works, first emerged in America. This gave the U.S. a huge head start over the rest of the world – including China, putting the U.S. firmly in the lead of the race for AI.
What Americans didn’t develop at home, they bought from Europe. In this respect, two British firms stand out with groundbreaking contributions to AI development: ARM and DeepMind.
While all eyes are trained on the AI race between China and America, is there a role left for Europe?
From the start of the digital revolution, and in spite of America’s lead, Europe has always had a fundamental role in digital research, a role often overlooked and even downplayed by the media mesmerized by Silicon Valley fireworks.
But the fireworks are dying down and getting messy now while China is on the rise.
America’s AI Roots in Europe
Let’s take a closer look at ARM and DeepMind, the two British firms that played a fundamental role in sustaining America’s lead in electronics.
ARM, a venerable British chip designer since the 1980s, was the first to develop a new architecture for processors – that’s what makes computers work – far more potent than anyone else. That system was bought by the American giant Qualcomm (also founded in the 1980s) and today the largest mobile chipmaker in the world. Since 2007, its ARM-based Snapdragon suite of system on chips (SOCs) has spread relentlessly and is now found in most smartphones.
DeepMind Technologies, founded in 2010, is a developer of artificial neural networks acquired by Google’s Alphabet in 2014. DeepMind is at the forefront of deep learning, a revolution in AI development that happened quite recently, in 2012 as a result of the brilliant work of scientists like Yoshua Bengio, Geoffrey Hinton and Yann LeCun who were given the prestigious Turing Award in March 2019.
It was an extraordinary conceptual and engineering breakthrough that marked a major departure from how computers had been working up to that point. Instead of relying on algorithms for specific tasks, deep learning uses a different architecture that imitates the way our mind works and permits “machine learning” in the way humans “learn”. To do this, they require to be fed vast amounts of data. The more data they get, the better they detect patterns in the data and “learn”. Learn to recognize faces. Learn to identify winning strategies in games.
DeepMind made headlines in 2016 after its AlphaGo program beat a human professional, Lee Sedol, the world champion in Go, an ancient Chinese board game that is far more complex than chess:
Competing on the edge of AI development: the U.S.-China 5G War
5G is a telecommunications standard that will allow for faster transfer of data in mobile networks. To understand what the fuss around 5G is about, we need to look at the political implications of the 5G spat between China and the U.S. who are, of course, both involved in developing 5G. The major player on the Chinese side is the tech giant Huawei.
Much more than a “spat”, this has all the trappings of a trade war. Trump has gone all out against the Chinese version of 5G. In one word: This is an area where the U.S. is bent on remaining “America First”.
What the U.S. has done is declare “war on Huawei”, to use economist Jeffrey Sachs’s term in his December 2018 article. The war has included some brazen acts, like the arrest of Huawei CFO Meng Wanzhou in the Vancouver airport en route to Mexico from Hong Kong, and then demanding that Canada extradite her to the US.
Canada caved in, Huawei’s CFO was extradited, but overall results so far are not encouraging for the U.S. camp. True, the uproar in February 2019 over the Thai government’s agreement to allow Huawei to participate among a range of companies in a country-wide test bed for 5G technology is symptomatic. Suddenly it looked like Southeast Asia had become the latest battleground in the U.S.-China 5G war. But the 5G test bed went ahead anyway.
Elsewhere reactions have been muted or slow. The Philippines has decided to work with Huawei while Malaysia and Australia are taking time to decide, as is most of Europe, though some, like Monaco have decided to go ahead with Huawei.
Also of note: The 5G war is waged against a backdrop of turmoil in the market for computer chips (also known as semiconductors). Major players are affected, in particular Qualcomm.
In a weakened state today since its peak performance in 2014, Qualcomm has become a target for acquisition. The Trump administration blocked the Singapore-based Broadcom purchase citing security reasons because of its partnerships and investments in China, in particular Huawei. And now rumors are swirling about other possible takeovers of Qualcomm, including by Intel.
There are other firms with an important role in AI on the market for acquisitions, notably California-based NVIDIA and NXP, a Dutch global semiconductor manufacturer. NVIDIA is the leading producer of graphics processing units (GPUs) which power today’s top video games.
The point about NVIDIA is that its GPUs provide a solid foundation for more complex AI applications (such as face recognition). NXP, in contrast, produces cheaper, simpler chips connecting devices in a secure way, particularly automobiles. In NXP’s own words, it “is driving Internet of Things (IoT) innovation in the secure connected vehicle, smart connected solutions, and end-to-end security and privacy markets”.
Both companies are dealing with the fallout from a Chinese economic slowdown, caused in part by the U.S.-China trade negotiations. The ongoing fall in the auto market has also affected NXP sales and NVIDIA is facing further headwinds caused by the cryptocurrency crash of 2018. Because GPUs are used heavily in cryptocurrency mining, NVIDIA had a huge amount of unsold GPU inventory at the end of 2018 and it saw its gaming revenue cut in half.
Just watch what would happen if China tried to buy it. But China has embarked on another course of action.
China’s Bid for AI Leadership
For China, watching its national Go champion lose to Google’s AlphaGo was its “Sputnik moment”. All of a sudden, China realized it could not let this happen and an all-out race with the West was on.
The government launched a vast plan to become the world leader in AI, announcing it would dominate the industry by 2030. It’s not just a matter of throwing tons of money at the problem (though that’s what China does). It’s the way the Chinese government is doing it that is especially effective. Every level of government down to city mayors is involved. Cities are spurred on to develop their own local Chinese Silicon Valley, with investment support, tax rebates and a myriad of incentives.
To Western observers, it may look like the Chinese government is throwing money unwisely, with many people taking unscrupulous advantage by rebranding themselves as AI startups. But brute forcing development has worked well for China before – for example, with high speed trains – and it may very well work again.
In any case the stakes are high. PricewaterhouseCoopers estimated in 2017 that AI deployment will add $15.7 trillion to global GDP by 2030, and that China will take $7 trillion of that total while North America will take far less, $3.7 trillion and Europe comes in third place, with $2.5 trillion, and “developed Asia” (mainly Japan, North Korea, India, Taiwan, Malaysia, Philippines, Indonesia) far behind, with $0.9 trillion.
That result, seeing China as the winner in the race, is something Kai-Fu Lee, a Taiwan-born (1961) former Microsoft scientist (he is the author of AI breakthroughs) and President of Google Asia, would agree with. Since 2009, he has turned venture capitalist and finances Chinese startups from his office in Beijing. As a result, he has a front row seat to witness the unfolding of the AI race from the Chinese side.
His book, AI Superpowers: China, Silicon Valley, and the New World Order, published in September 2018, is full of deep insights into the Chinese digital world. Unsurprisingly, it was an instant bestseller. He details how the US and China are driving a deep learning revolution and why China is likely to be the winner.
The reasons why are simple enough: demographics and big data. China, by sheer numbers, beats all the other markets. And, as Kai-Fu Lee puts it, China is the “Saudi Arabia of data”.
This, he argues, is a result of Chinese digital firms’ “heavy approach” to the market, systematically using O2O (online-to-offline) strategies, whereby firms are not afraid to “dirty their hands in the real world” and follow through with services right to the consumer’s doorstep. This is in contrast to the “light approach” adopted by American firms that essentially stop at coding and do not step out of their offices.
As a result, Chinese firms gather far more data about the real world than Amazon (that relies on vendors to ship goods) or Facebook. This, argues Kai-Fu Lee, should give China a head start in deep learning: A machine “learns” from the data you provide them with. The more data it is fed, the better it gets.
Take a minute to watch this video, the end, with Kai-Fu Lee’s recount of a deeply emotional near-death experience is fascinating. Note how Kai-Fu Lee is remarkably optimistic about AI: He sees automation and job replacement not as a threat about to disrupt our lives but as a genuine opportunity. “AI is serendipity,” he says. “It is here to liberate us from routine jobs, and it is here to remind us what it is that makes us human.”
The other thing he seems not to fear is the possibility that AI will invade our privacy. That AI devices might manipulate us to their own ends (or to an AI master’s ends) – possibly without our even realizing it.
In short, AI ethics is not something he worries about – and that may be a specifically Chinese cultural trait. Kai-Fu Lee admits as much when he writes: “People in China are more accepting of having their faces, voices, and shopping choices captured and digitized.” And he adds: “This is another example of the broader Chinese willingness to trade some degree of privacy for convenience.”
Convenience? This is something European and American consumers may not feel so comfortable with. And surveillance is everywhere in China. As Kai-Fu Lee notes, “That surveillance filters up from individual users to entire urban environments.” He does note in his book that Europe has taken a much stricter approach, as evidenced by its adoption of the General Data Protection Regulation (GDPR) a law to protect user privacy that has no equivalent in either China or the U.S.
Europe’s Role: Renewed Focus on AI Ethics
China’s one party system and Communist philosophy no doubt contribute to an optimistic approach to AI, while in America the government has basically kept to a hands-off position, leaving it to the private sector..
In America, so far, the worry about AI is private, in Europe it has gone public. And it goes well beyond GDPR.
Consider the facts. In America, Elon Musk is the one tech billionaire that has played the largest role in raising awareness of AI threats to ethics.
Musk began to worry about AI ethics several years ago, speculating that AI was humanity’s “biggest existential threat”. His alarming talk at MIT in 2014 went viral. That year he donated $10 million to the Future of Life Institute for research to ensure AI is used for good, not evil.
In 2015, Musk announced the creation of OpenAI, a $1 billion non-profit foundation. Based in San Francisco, OpenAI aimed at promoting safe artificial intelligence research and by 2018, it had acquired a range of major supporters. Soon most big tech moguls joined Musk in expressing concern, notably Bill Gates.
But the Federal government refrained. President Obama announced a plan to support AI development in 2016 without any funding specified. As that plan came out during the presidential campaign, it went largely unnoticed.
Now Trump, after two years in power, has finally awoken to the need to support AI development. In February 2019 he signed an Executive Order building on Obama’s effort and enjoining federal agencies to invest more in artificial-intelligence R&D, share data and computer models with outside researchers, establish clear technical standards and boost workforce development – but still leaving the bulk of the effort in private hands.
Bloomberg editorial board considered it a “smart plan” and a good start, though it acknowledged it was not enough compared to the billions China is pouring into AI deployment.
And it is certainly insignificant compared to what Europe is planning to do.
Three days ago, on 8 April 2019, the European Commission presented a set of guidelines to ensure AI is developed and deployed in an ethical, fair way. The guidelines, after a period of testing with the public will come on stream by 2020. They are just one of the “deliverables” under the EU AI strategy of April 2018 adopted by Brussels, which aims at increasing public and private investments to at least €20 billion annually over the next decade.
That’s some €200 billion, a huge amount.
But it’s not just a question of the amount of money set aside to develop AI. It’s a question of how to develop it. The AI ethics guidelines may be just one “deliverable” and to some, they may look unimportant, but they are a fundamental first step, an opening salvo that sets the stage for AI development.
On 9 April, there was a press conference outlining EU Commission goals in AI, with Mariya Gabriel, Commissioner for Digital Economy and Society, who focuses on AI and digitizing the European cultural heritage and Commissioner for Agriculture and Rural Development , Phil Hogan, focused on farmers:
The language in this video may sound somewhat bureaucratic and it is unlikely to go viral.
Yet make no mistake.
This is the only serious effort by any public sector in the West. It is far larger than what any single European country is undertaking – for example, Macron has set aside just €1.5 billion fo AI development in France over 5 years.
The EU effort to support AI is multiple times anything a single European country can come up with. It may not match the Chinese effort but at least it stands a chance to make a dent.
Like the Chinese, one can expect that the European investment will result in making more data available, fostering digital talent and supporting AI breakthroughs, helping to spread adoption of the new technology.
Yet it will do something that the Chinese haven’t even considered: It will try to build trust. It addresses the AI threat to our lifestyles, jobs and values head on.
That is Europe’s bet: Money is not enough; what is needed is a trustworthy and fair environment for AI to flourish in. And to save humanity from the “AI apocalypse” feared by Elon Musk and Stephen Hawking.
Humanity, however, won’t be saved from AI threats if the populist parties rising everywhere in Europe have their way and win big at the coming European Parliament elections in May. Should they succeed in dismantling European institutions as they so dearly wish, you can kiss goodbye to a fair and ethical AI shaped by Europe.
The reason is simple: no single European country can do the job the EU Commission can. As a result, AI will be shaped by the U.S. private sector and China.
But mostly by China.