Why China May Be the Next AI Superpower

Lassi A Liikkanen
5 min readNov 13, 2018

This is a recap and a review of Kai-Fu Lee’s book AI Superpowers.

Picture of Kai-Fu Lee (2018) AI Superpowers

Part I — Preparations for a power struggle

Most technologists have observed over the past five years how a new generation of Chinese technology entrepreneurs have entered the global spotlight. At least Jack Ma of Alibaba and Lei Jun of Xiaomi have become familiar faces also for Western audiences, head to head with Silicon Valley superstars. However, these visible Asian businessmen aren’t even together as popular as Kai-Fu Lee in China. Lee currently racks a staggering 51 million followers on the Chinese instant messaging platform Weibo. For comparison, Bill Gates has 46 million on Twitter. To arrive where he is today, Lee has been the leading researcher in speech recognition in the 80’s, a top manager at Apple and Microsoft, and most recently a venture capitalist in China. So when this Taiwanese technologist has a word to say about AI, China and the social impact of entrepreneurship, at least I’m willing to lend an ear.

Photo: Kai-Ful Lee on TED stage in Vancouver, August 2018, first appearing on TED stage in 1992 for achievements on Apple speech recognition. YouTube Screenshot

What does it take to rule with AI?

Lee’s visions of the future world order rest on few premises he lays out in the beginning. First comes the claim that deep learning is a key technology has made a huge improvement on the capability of AI applications. The impact is so big that it is unlikely going to be followed with anything similar in the near future. Thus, most companies can now reap returns by implementing the latest AI technologies instead of investing in research for new technical discoveries.

To successfully deploy AI, you need to align four assets or forces:

1. Computing power

2. Data

3. Data science expertise

4. Policy

According to Lee, China is right now in 2018, perfectly aligned to deploy AI widely across its vast technology industries. As the author quite critically assesses China’s position, it becomes evident that China has, in fact, already assembled an excellent combination of these assets.

China leads the era of implementation

Affordable computing power is ubiquitous nowadays. Money has also become a non-issue in China since private and governmental investments in technology, particularly in AI, have become very generous. But data is something China has an ultimate advantage in comparison to anywhere else in the world.

The rocket growth of Chinese mobile internet infrastructure has created an unimaginable pool of data in both quantity and quality. Quantity follows naturally as the great nation has adopted mobile services, but quality is a consequence of technology integrating far deeper into everyday life than anywhere else in the world. Quality is also an effect of Chinese policy. Without privacy concerns such as those embodied by European GDPR, the digital footprint of a Chinese citizen is a rich asset for AI technologies. For instance, the digital rent bike network alone creates a huge amount (20TB/day) of detailed citizen travel data and digital wallet enable tracking all of their financial transactions.

The factor where Chinese still are behind the current leading AI superpower, the United States, is engineering excellence. The brightest minds around the world still converge at the biggest American companies and their offices in California, Washington and Massachusetts. But Lee believes that as engineering education in China has taken huge leaps in the past decades (and he has participated in this), China is now prepared to feed the AI industry with an increasing workforce that is capable to successfully deploying the latest American ideas. Ideas, which according to the tradition of data science community, are shared for free in real time.

The execution in big numbers matters more than individual breakthroughs, Lee believes. He does have mention a disclaimer: if the next deep learning comes about in a private lab of a tech giant, it might tip the scales for very long time in favor of the lucky breakthrough makers. What he doesn’t say is that by early 2018, according FHI study the Chinese pool of AI “researchers” was still only half of the US equivalent (39,000 versus 78,000), so it will take years before they are even on quantitative scale.

Can public policy make a country innovative?

Lee spends a better part of his book’s first 100 pages dispelling or clarifying the copycat culture that Chinese “design” has become known for (I’ve also explored this topic this six years back). In short, it is claimed that the replication of western designs was still a viable business ten years ago, but that is quickly becoming the past. First Chinese used Western design as a basis for leapfrogging to unique products, then came the Chinese innovations. Now we are seriously looking at an era where Western companies may start creating local replicas of Chinese origin!

Screenshot of Kai-Fu Lee’s presentation at O’Reilly AI Conference 2018 https://www.oreilly.com/ideas/china-ai-superpower

Chinese innovations are already penetrating to international market on their own. Anyone looking at the global smartphone market shares can’t deny that with Huawei, Xiaomi and OPPO, Chinese companies are original enough to compete with the best of rest, and not only among Chinese customers.

The main argument is that the economical triumph of China over the past decades was achieved with a clear engineering mindset and practical support for building things as fast as possible while letting the market dynamics sort out the successful ideas. In its current state, the Chinese officials are pushing for yet another “Brute-force economic transformation”. This time they expect to do it using AI. And governmental forces are willing to push far, incredibly far by Western standards, to achieve this. For example, Baidu is collaborating with local government to basically develop a whole city with target population of over two million to make a proper test ground for autonomous traffic!

Xiong’an AI City on YouTube: https://www.youtube.com/watch?v=RzO22w9L-aA

The net governmental investments and subsidies in technologies, partially in AI, run in billions of dollars annually. And it is not a Silicon Valley of China that the country is building. The development funds are distributed all over the one billion inhabitant country! In 2017, China announced an ambitious mass entrepreneurship and mass innovation program by which it aims to accelerate innovations from startups. The Americans who still lead the semiconductor business (the material brains of a computer and smartphone alike) may be anxious to hear that the “Made in China 2025” project aims to take the lead in chip development by 2030. Lee doesn’t believe all ambitious goals and development plans will hit home, but overall impact will notable even if only 50% would go to waste.


Part II published on 20th Nov 2018 will further discuss Lee’s view of Chinese tech company scene, reveal Lee’s plan to save humanity as well as reveal few notable omissions in his treatise.

Fourkind is an emerging technologies consultancy helping ambitious companies world wide to attain a strategic edge



Lassi A Liikkanen

Data loving designer & inter-disciplinary researcher interested in technical innovations, design creativity and about how emerging technologies affect CX.