China's entrance as an AI superpower is a thriller economics. No economy ever changed on such a grand-scale so fast in bleeding edge technology. China began primitive in AI ecosystem and ended up as world leader all in a decade. A recent Wall Street Journal article says, as of 2016, China had invested $46 billion in various Silicon Valley startups. Almost every other AI startup in the world is funded by Chinese capital. China's this change shames even the best of corporate change history be it IBM learning to dance or GE shedding fat. Above all, it proves an inspiring point. With a war strategy to outsmart and rigor in pursuit - even a large economy can master rocket science in just over a decade. That should be great news for the c-suite of billion dollar legacy enterprises. Many of them out-maneuvered by Silicon Valley startups - are now injecting digital into their bloodstream.
How did China do it all so fast? How did a country with 90 percent of its population below poverty a few decades ago, a country that had every conceivable economic problem - turned everything on its head and sat on top of the supreme tech game in the world?
Many leading AI researchers like Kaifu Lee have a theory to explain this phenomenon. What escapes every liberal economist, and management guru alike in search of such a theory is - the surprise swung by the Chinese government.
China didn't just throw a bunch of liberated entrepreneurs and waited for an AI miracle fingers crossed. Nor did it open-up freely its economy to American Gods of AI. It crafted cleverer policies and adopted sharper practices. Sure, a lot of it were learnt with experience and evolved overtime yet, policies were more premeditated strategies than piecemeal ensemble of disjointed rules.
Government played a supporting role
China played the godfather for its AI startups. Anyone who has had a slight brush with bureaucracy knows, government playing the benefactor can turn soon sinister at the point of action. Well intended policies can entangle the entrepreneur in a bureaucratic net of approvals, clearances, of access to key resources and so on.
China is the AI superpower that is at the edge of overpowering the US supremacy, says Kaifu Lee. Lee is one of the leading exponents of AI in the world. He saw AI evolve very closely as a student of scientist Dr.Raj Reddy, and then as head of Google operations in China. Lee later left Google to found his own venture capital for Chinese AI startups. He is not the only one. Research by MIT Technology Review showed that China published more and also wrote higher quality papers in world leading scientific journals such as Nature etc in the last five years. Chinese research talent is writing a whole lot of cutting edge algorithms.
China unlike the US didn't come up the AI ladder the linear way. In the US, AI was logical extension of emergence of e-platforms and smartphone revolution which followed widespread adoption of software, laptops, desktops and some form of mainframe industrial computing. China from industrial computing went straight to smart phones and mobile commerce to platform economy. Chinese consumer was too poor to buy the desktop and laptops. Most Chinese users didn't transition from laptop to smart phone. He started with smartphone straight. This had, selective disadvantage as classic strategy guru Dr.Michael Porter puts it. This disadvantage made China look at situations differently. Sometimes certain victory can confine the thinking. For instance Microsoft had lost smartphone, mobile platform and cloud due to its glorious victory at desktop operating system. Not having such obsession was advantage for Amazon. Similarly China missed whole wave of this desktop and laptop computing and that became a huge advantage when it came to adopting mobile technology in a way. The entire energy that was not invested in that, now was available for adopting smartphones. Whole generation of hungry for success entrepreneurs plunged into this new field called artificial intelligence.
It is different with Chinese
Well there is an interesting twist to this Chinese AI entrepreneurship. Conventional wisdom says that innovation is bottom up, it is driven by free market forces. For AI in China it was reverse. China's AI innovations and widespread adoption was driven top down. It was initiated majorly by the government says Zen Soo one of China's top tech researchers and journalists. Kaifu Lee agrees, he says Chinese government played a powerful role. It established critical ecosystem that helped the AI startups spring up and thrive. First it provided start-ups the physical infrastructure. And second, it also gave easy access to large funds. Government specially designated physical infrastructure for hi tech startups which had internet bandwidth and other pre approved facilities. Hundreds of cities in China had all these facilities and these were made available at highly subsidized costs. Also government simplified and quickened the dragonic processes of availing these facilities.
Government of China itself created huge funds for investing in AI startups says Kaifu Lee in his book - AI superpowers. The government formed agencies and invested billions of USD in risky ventures through private VC funds as their partner sharing larger risk in case of loss and making it doubly profitable in case of success. This was one of the most innovative economics ever tried. In their book on Secrets of software success, authors Detlev Hoch and Cyriac trace early American government funding for software in the form of large software development contracts to private software service providers. In his massive research Harvard strategy professor Michael Porter found that governments across the world played a key role by way of giving funds and shaping policies in building competitive advantage in select industries. Chinese government invested billions of USD even directly, partnering with private venture funds in a massive way.
Another part of this economic metamorphosis is the role of former Chinese students from American Ivy League universities like Stanford, MIT etc where the AI technology is brewed. Many current day China AI fast growth have US Ivy League connection. As China opened its economy for private capital. Many such Ivy League Chinese toppers came back to their country and started dabbling with IT ecosystem. For example meitu, bitman, mobike, etc have 100 mio USD valuation. They were all started by former stanford and other Ivy League students from China. Lee himself is a PhD from Carnegie Mellon University and was a part of Microsoft and Google at senior level. His VC firm Sinovation ventures funds the early AI startup in China and has a portfolio of billion USD plus . Wang Xing serial copycat is a student from the University of Delaware. Wang took some of American hottest startups like Facebook and made xiaonei for Chinese users.
Chinese VCs in Silicon Valley
Yet another key part of the success story is Chinese venture capital in Silicon Valley startups. Today, Chinese private capital is as liberated as any other in the world and many have started investing in cutting edge tech startup around the globe, especially in silicon Valley. In fact China is one of the top investors in the Silicon Valley ventures. It appears that long history of communism has doubled its appetite for AI venture capital. In 2018 Chinese silicon Valley tech investment peaked at USD 3 billion. Once such investments crossed the threshold these board level connections gave key insight into technology that could be customized and implemented in Chinese ecosystems. Kaifu Lee, in his book says 48% of the world’s funding in AI startups came from China in 2017.
Also three big China tech companies alone accounted for 44% of equity investments in US AI startups between 2014–2018. In five years every other billion dollar AI company in the world would be making money for China, making it and unconquerable AI Gladiator. That's not a mere entertaining hypothesis at the water cooler. For instance, in India, Flipkart, Byjus, Ola, Paytm, TikTok, Swiggy, etc have been hugely successful and are mainly controlled by Chinese funds. Though founded by an Indian, large controlling stake, about 40% at Paytm is with China. According to an article in The Hindu, China invested about 5.6 billion dollars as of 2018 in many Indian startups. Alibaba, Shunwei Capital, Fosun Tencent and Xiaomi were among the first to invest in India and lot of them are pure AI or essentially AI driven. With all these investments China has distinctive advantage over Indian data and Indian talent. According to some experts, these startups have attracted the best of Indian AI talents as against the India Google and big four Indian IT employers, Infosys, TCS etc.
Most large scale AI engines - AI that can transcribe Hindi or Chinese, recognize face for the bank enquiry, read the chest x-ray, detect a cancer etc - have three parts. First part is the essence, the know-how, the algorithms behind the AI. This is the original newness in the software, innovation in the methodology, different models of AI.The second part of the AI is the data. Data plays a huge role in making AI useful. Without the data the algorithm alone is toothless. In a crude sense, algorithms are like spreadsheets. Without the data spreadsheet is of no use. It has capabilities to analyse, do maths, filter tables, and draw the charts. But without data what charts will you draw what tables can one make. Data makes the spreadsheet useful. Same is true for the AI algorithms, without data they are not of much use. Third part is the cost, feasibility of computing. The physical infrastructure of computing, the processor, the internet bandwidth, the hard drive to be precise. The lesser the cost of computing the more affordable its adoption.
When it comes to part one, algorithms and new methods, the US has a thin advantage over China. Research and academic infrastructure in the United States are definitely far superior. AI researchers in the US are lot more in number and slightly better in quality. Though China is fast closing this gap, as of now US leads this part. Interestingly enough China has not invested much in original research. Most of the algorithms used in today's AI world were discovered in the US universities. For instance the Holy Grail of AI, deep neural net was first discovered in a Canadian university. Canadian government had funded this research for very long. These funds helped British professor Geoffrey Hinton, currently director at Google AI discovered practical deep neural net that could recognize speech. According to many research reports available the research was carried on over two decades, about a hundred million or so was spent on writing many unsuccessful AI codes and collecting data.
It is in part two and three, in data and cost of computing China has gained unsurpassable advantage over the US. This partly also explains the apprehension of the the US journalists hi tech press.
As explained earlier the AI advantage, for organisation or economy comes from the AI know-how the methodology, algorithm and models. However, thanks to America's global economy and open academic University environment, China has the same access to AI Technology as American companies. Because of free access America has no special advantage over these than China.
AI engine, especially the deep neural net, the poster-boy of AI innovation, requires a lot of data to learn and become practically applicable. For instance, Chinese companies developed very early in the game, face recognition system. This was made possible by easy access to millions of face images of people and use of such technology. Government was first to use these services. This early access to profitable customer gave Chinese firms big advantage. In the US and Europe privacy laws, culture and sheer access hindered AI companies attempting training the face recognition engine, while in China because of the Government support people could easily access face data a large amount of a state was used to train a engine that became more and more accurate. Faceplusplus for instance was one of the first companies which had the first mover advantage over all the other companies in face recognition system. Megvii, Beijing based startup was entrenched in banks, trains, and offices in just two years. Another start-up, faceapp like AI engine that helps photo editing and beautification - has about a billion users prospered on a similar advantage of China's policy. In terms of algorithms and computing machine the west also has same resources but without the data those would be of no use. Take another startup in k12 space,which clearly owes its success to China government and the policy. Squirrel which teaches primary school students is valued at 1 billion US dollar in little about 5 years. According to an article in MIT Technology review it has 20 million students being trained on its AI driven platform at 3000 plus training centers across China. About a billion usd has been invested so far. This was possible only in China. The government was more non- interfering in deploying such AI in traditionally very controlled area like primary teaching. Eminent educationists and researchers in the West have been talking about it for very long but they were restrained by the convention.
There is also an emotional factor playing in favor of China, according to Lee. Chinese entrepreneur is more hungry for success than his US counterparts. They compete far more fiercely than his Silicon Valley opponents. Chinese startup founder carries on his back, the generation of dreams of his grandparents who have lived in poverty all their lives. He is not just a businessman, he's a gladiator. As against this Silicon Valley is driven by passion, innovation and led by rich kids with safety nets in case of failure.
-------
References
Book by Kaifu Lee - AI Superpowers - https://www.amazon.in/AI-Superpowers-China-Silicon-Valley/dp/132854639X
https://www.wsj.com/articles/chinese-cash-is-suddenly-toxic-in-silicon-valley-following-u-s-pressure-campaign-11560263302
https://www.scmp.com/business/companies/article/2152935/hangzhou-chinas-answer-silicon-valley-hit-returning-graduates
Book - https://www.amazon.in/Secrets-Software-Success-Management-Insights/dp/1578511054
Book -https://www.amazon.in/Competitive-Advantage-Nations-Michael-Porter/dp/0684841479
