Review of AI Superpowers
I got my hands on the Artificial Intelligence(AI) Super-Powers By Kai-Fu Lee when I received this book as a birthday present. As an aspiring computer scientist and an AI enthusiastic, I would be remiss if I didn’t read this book.
Lee starts the book by comparing Silicon Valley to China’s “copycat” companies. He mentions how many at Silicon Valley views them as mere copies of the prestigious valley products. However, he further explains how incorrect that notion is. Due to the lack of regulation by the government and these companies’ hunger for money, they are willing to do anything, copy any model, or follow any business that can achieve these goals. Initially, these companies are copied, but with a market rooted in competition, they have to and continue to innovate and keep themselves planted. An example of that is WeChat, which was initially modeled off WhatsApp but innovated to incorporate voice messages, online banking, and many features that make the app a seamless ecosystem.
From smart replies in Gmail to apps that provide directions, it is everywhere. Lee mentioned an anecdote from when he visited a Beijing kindergarten classroom and was surrounded by curious mouths. As they questioned him from silly questions to serious ones like asking, “If robots do everything, then what are we going to do?” These questions remind us how unsure we all are of the future of A.I. and provide insight about A.I. in China.
A few years ago, it seemed that China was lagging years behind the U.S. in artificial intelligence. In a matter of a few years, China has become the only rival of America when it comes to emerging technology.
But where does this A.I. journey even begin? It started in the 1950s when the developers of A.I. were attempting to recreate “human intelligence into a machine.” They were two main approaches, one “rule-based” and the other “neural networks.” Rule-based gave the machine rules that if X, then Y, while this could work for things with small possibilities, but what 200, 1000, or even 1,000,000 possibilities. However, neural networks attempt to create connections in the machine similar to our brain, rather than telling, teaching the computer through trial and error. Like the human brain, computers make connections and teach themselves to identify patterns within data.
This innovation lagged until the development of smartphones that allowed for the discovery of a way to train new layers in neural networks, which is known as the “deep learning” revolution. This brought an ocean of potential for A.I., where technology could recognize patterns, optimize decisions, and solve problems, it could think. This created a boom in many nations like the U.S., Canada, and the U.K. but China wasn’t part of this race until its “Sputnik Moment” in 2016.
Like the launch of the first human-made satellite into orbit by the Soviet Union pushed the American government to take action. It brought fear and realization about how behind we were in the race of space exploration. This pushed the government to create NASA and begin funding for math and science education, creating an avalanche of innovation and allowing them to beat the Soviet Union. Similarly, in March 2016, when Lee Sedol played Go(an abstract strategy board game played on a 19x19 grid amongst two players where each aims to surround more territory than the other) against AlphaGo(an A.I.), and the A.I. won four-to-one against the legendary champion. It became the catalyst of Chinese technological advancement. Although this event wasn’t watched by many Americans, more than 280 million Chinese viewers had watched this moment and realized their lag in this field.
While Silicon Valley ridiculed the “copycat” notion of Chinese companies, it didn’t realize that these companies were learning how to design products and replicate them pixel by pixel and improve them in the increasingly cutthroat Chinese market. Lee mentions many Silicon Valley companies transition into the Chinese market and failing because of their inability to adapt, even Google. Lee suggested that there was a spark difference between Chinese and American users on Google, where American users spent few seconds in the search tab before clicking away, Chinese viewers spend 30–40 seconds looking at a couple of links at a time. Lee suggested that Google should adapt to the Chinese market by allowing users to see the search results and website outlook on the same page, but Google’s inability to adapt kicked it out of the market.
Zhongguancun, China is a technology hub, similar to Silicon Valley but unlike Silicon Valley, where a product is trying to fill the gaps like providing a service that connects drivers to riders through an app like Uber. China’s goal is to control the whole process from top to bottom. They want to hire drivers, own gas stations, and repair shops to keep the app going as Didi. Lee notes this as the O2O revolution, which means online to offline. Because these startups were created on the idea of copying, they want to ensure that their product can’t be duplicated.
To further this advancement and a major player in China’s growth was the Chinese government’s mass innovation campaign, which provided an incentive to start tech startups and made the Chinese people trust them even more since they were backed by the government. Chinese technological culture differed from Silicon Valley, where the Valley is mission-driven and Chinese companies are market-driven. For Chinese companies, the most important thing is to make money; they are willing to make any change to ensure that the market is happy, so the money keeps rolling, while the Valley focuses on their image and their mission.
Four Waves of Artificial Intelligence:
They are four waves of A.I. that are affecting our lives. The first is internet A.I. through the recommendation engine. Over the years, this engine has improved to recommend the users better products, increasing the probability of purchase. This is directly based on data from the user base, and China has an advantage over the U.S., because its internet usage is greater than the U.S. and Europe combined due to its large population and a population that had full trust in the internet, which promotes innovation.
Next is business A.I. which is A.I. that allows a business to gain insight about their users through data analysis and engaging with customers and employees. In this, the U.S. has an advantage simply due to the large data sets that cover multiple decades, while China doesn’t have a centralized data system and outsources its data, which makes this type of A.I. harder. However, there is one exception, Smart Finance: since China skipped the credit card revolution and went to straight to online payments, they do the same for personal loans. Unlike banks that look at financial history, it gives off loans based on how quickly you type in your date of birth and how much battery power the device has to make the loan decision. This is based on other users with similar results and Smart Finance’s experience with them, which has proven to be reliable. Lee believes that the U.S. has a 90–10 advantage, but in some time, China will be able to catch up through the fields of medicine and law. Due to the large number of China’s population, there aren’t enough doctors to the number of patients, and there aren’t enough judges to make a fair decision, which makes having A.I. helpers a necessity furthering innovation.
Next up is perception A.I. which deals with the computer’s ability to recognize objects and sounds, like the human brain. For example, when you walk into a grocery store, the cart gives a list of things that are low in the fridge, your food preferences, and as you add things to the cart, it is deducted from your WeChat wallet. This helps with the merge between online and offline. This advancement exists in Shenzhen, China, with the Xiaomi line similar to Amazon’s Echo, which also includes AI-enhanced rice cookers, refrigerators, speakers, and even vacuum cleaners at an affordable price, which gives China a 60–40 advancement.
Lastly, autonomous A.I., which is the combination of the first three. These places robots in place of humans, where he predicts there will be drones to paint houses, fight forest fires, and could possibly provide disaster relief. However, one reality of this A.I. is self-driving cars, where American companies like Google and Tesla provide a 90–10 advantage. However, China building highways specifically designed for A.I. vehicles, equipped with tools that allow your car to see. He predicts that in close to five years, China and U.S. will have a 50–50 advantage.
With AI becoming so powerful, many fear that robots might take over humans, which isn’t a realistic fear. However, one that is, A.I.’s take over multiple sectors of jobs. Where many studies predict 47 to 38 percent of the risk of job loss due to A.I. Further, the studies don’t take into account industries that could be fully disrupted, like banks with the earlier example of Smart Finance.
Shift in View
However, Lee had a similar view until he was diagnosed with stage IV lymphoma, where he was grateful to A.I. for its ability to detect and diagnose his condition, but what brought him back was the support of his family and community. It also brought the realization that the love and care that humans can deliver is not possible by robots. Like, A.I. could take over the medical profession, but only humans can provide consoling and emotional support. He proposes a radical approach and suggests the government to provide an incentive to care, whether it is to an older parent or a stranger, to reward doing what we do best that is “loving and being loved.” To change from an economic driven to a socially productive one.
From the 1950s, artificial intelligence was explored, and its use was furthered by the invention of the smartphone. China entered this innovation and added to the technological revolution and became an A.I. superpower, by the help of its government. China follows brute force approach when it comes to deep learning models, as it supposedly works better for larger and less experienced population of Chinese software engineers compared to the U.S. As mentioned earlier, about the upper-hand of U.S. in the collection of data. Lee believes that due to China having few data protection regulation, this allows Chinese software to collect more data from their users, which will eventually provide them an upper-hand. I wonder if a large amount of data but only over the course of few years can compete with data that expands over decades. Further, the author doesn’t account for the fact that China has one of the worst track records of human right violation. With people in the U.S., realizing how manipulative A.I. is and the government considering regulating the amount of data companies can keep. I can’t help but wonder if China would ever curb to this and if the people in China are having the realization that they are selling their convenience for a small portion of themselves. As Lee mentions about A.I. and humans working together to make the world a better place, another question that comes to mind, is all Chinese A.I. attempting to make the world better. Recently, I heard of China’s Huawei testing an A.I. whose sole purpose is to recognize Uighur Muslims. Lee states that China and the U.S. will be split 50–50 in autonomous A.I. due to China focusing on building A.I. equipped highways, I find it hard to believe due China’s multiple construction fails. As A.I. continues to innovate and there is a fear of job losses. However, Lee believes there is a need of a shift in our minds where our societies become more socially-focused and reward compassion and the human condition, which allow for the existence of A.I. and humans.