In advancing AI, there is competition between the USA and China. It is predicted that AI will possess the technological skill to remove 40 percent to 50 percent of occupations in the USA. This prediction is terrifying, however, it is significant to highlight the “technological skill” condition. Currently, China has important developments to become the leader of the industry. Within ten years, China will suppress the USA in leading Artificial Intelligence worldwide, pushing smaller nations to leave behind. Kai-Fu Lee explains China’s benefit in AI advancement exceeds the competitive edge.
Chapter 1 – Thanks to the advancement in deep learning and machine learning techniques, we are close to an AI economy.
Up until now, when people discuss Artificial Intelligence (AI), they were possibly talking about science-fiction movies. However, currently, everybody from school children to executives is considering the types of conditions AI will show us in the forthcoming days.
When the writer of this book gave speeches at schools and conferences that managers attend, he has observed that students were directing the same inquiries as managers: “Will teachers be replaced by AI?”, “Which jobs will come up in the future?”, or “Which jobs will be destroyed in the forthcoming years?”.
While AI is thought to be a current phenomenon, it has been around for decades but has just currently begun to be used as an important business instrument, owing to the advancement in deep learning and machine learning techniques.
The tale of how deep learning is found extends back to the 1950s, the times when scholars such as John McCarthy and Marvin Minsky were working on the aim of embedding computers with the human mind. In the early 80s, when the writer began being included in this field, two types of folks worked on this: the “rule-based” ones and the “neural network” ones.
Rule-based artificial intelligence folk thought that the best outcomes would be from programming computers with one rule each time, like “cats have triangular-shaped ears”. On the other hand, neural network folk wanted the computer to learn by itself, like people do, by experiencing. In this way, a computer can examine a photograph of a cat, and when it makes an error, it can learn from this data.
The thing neural-network-based folk wanted was a gigantic amount of data to examine and more rapid computing power, which eventually achieved in the mid-2000s. We now own more rapid chips and microprocessors. These better circumstances allowed AI researcher Geoffrey Hinton to put sufficient layers to the “neurons” and eventually multiply the AI processing power to another degree.
With this, the neural network started to be called “deep learning”. The huge discovery was announced at a 2012 competition when Hinton’s new AI algorithm has shaken the race in Visual Recognition.
Abruptly, AI could deal with complicated problems, understand patterns and find out great outcomes. It could be seen that this advancement could be implemented in a lot of activities, involving visual and audio recognition, speech interpretation, translation, complicated financial decisions, and even car driving. Owing to deep learning, the artificial intelligence economy is expected to be here.
Chapter 3 – China’s exceptional online presence provided a great opportunity for the type of data AI needs.
There are some big distinctions between Silicon Valley and China start-ups, and a major one is possessing something named light or heavy touch.
If a company has a light touch, it does only one thing and lets others handle the details around that thing. They provide some space for others. This is the manner of Silicon Valley start-ups such as Uber, a company that links people with rides, however, does not handle gas and car repairment.
The Chinese correspondent of Uber is Didi, which also possesses gas stations and repair points that make their rides perform well. This heavy touch design is liked in China because it is more challenging for an imitator start-up to completely copy the business.
A heavy touch and commanding every feature of a business would also enable more data, something that is very significant for a great AI output. Data is the food for AI databases. China already owns the biggest data treasure worldwide. This is specifically correct when one considers Tencent, the firm behind WeChat, an approvable super-app that everyone uses for anything.
To grasp the fact of WeChat, one should know that the majority of Chinese people are mobile-first Internet utilizers, in other words, their first internet utilization begins with a cheap smartphone instead of a computer. By considering this, WeChat has turned into a mobile application that allows people to do anything they would do with a computer.
The mini-applications within WeChat allowed not only messaging with friends but also enabled ordering food and grocery, bike-sharing, purchasing movie tickets, buying plane tickets, booking doctor’s appointments, ordering prescriptions, and securing stocks, everything within the WeChat app without opening or closing other applications or websites.
Plenty of these services are allowed by another mini-application: the WeChat Wallet, which is released on Chinese New Year of 2014. Each New Year’s Day, there is a tradition that people send money in red envelopes to their loved ones. WeChat made this possible electronically, without taking transaction fees from users, and they achieved great success with the release – five million people connected their bank accounts to WeChat and sent 16 million red envelopes by utilizing the app.
Following the utilization of the WeChat Wallet, Chine has turned into a raising cash-free community. That is a lot of data under one place, making it more bright what people desire to purchase, where they go and lots of other things. They analyze the data users generate.
Chapter 4 – China is the leading superpower in internet AI, however not in business AI.
The entrance of AI into our daily lives is occurring with four distinct waves.
The first wave is the internet AI, which has already arrived. YouTube offers videos for people to see according to an AI algorithm, and the Toutiao app offers articles plus produces them.
Considering the leader in internet AI, the writer thinks the US and China are close to each other for now, however, within five years he forecasts China will have a 60-40 benefit regarding the ability to leading the market. The fact that China has more internet utilizers than the US and Europe together, and a society that is ready for mobile purchases to content creators are the reasons for this forecast. Mobile applications such as WeChat Wallet already allow people to transfer micropayments like a few cents to web content creators they enjoy, and this kind of an atmosphere will allow innovative outputs from empowered creators – providing China a benefit.
The second wave comes with the business AI, which is the class in which the US has a big benefit. Business AI is currently arising with algorithms for decision-making in financial portfolios and bank loans. China currently possesses some fascinating mobile services such as Smart Finance which allows loans without looking at any financial history or zip code. Rather, it utilizes novel metrics such as the duration you take to answer some questions, and the battery power percentage your device has. With this, it became a trustworthy loan service for people who do not have a bank account or migrant workers and it turns out that the default percentage is very low with just single digits. Without any past information, it would require time to go along.
But a point that China is behind is business records. In contrast to China, the US has a flawless past of recording, with databases filled with banking, health, and other corporate transactions. This is why the US is in an excellent place for business AI and the writer gives a 90-10 benefit at this point. The forecast for the next five years is a little bit better for China, with the US benefit cut to 70-30.
Chapter 5 – China boosts game business in perception AI, however, the US has an early leading position in autonomous AI.
The third wave is the perception AI that involves voice and facial recognition algorithms. China has a benefit in this wave partly because of cultural distinctions. Americans have “Big Brother” hesitations when their image and voice are taken, on the other hand, the Chinese are more suitable to the idea of foregoing a bit of privacy in exchange for more comfort.
Perception AI is promising in being a thrilling wave as it erases the barriers between online and offline. This is the reason for this technology to be categorized as online-merge-offline (OMO).
An OMO implementation we will encounter more is smart grocery shops. Think of getting a grocery basket that scans your face, realizes you, and provides your shopping list. In this way, it welcomes you with the voice of your favorite actor. And because it scans all the items you put in your basket, it can warn you before you head to the checkout counter when you forgot to buy something. It could even recall you about your loved one’s favorite type of drink as you come close to that department in the store.
China is already producing the Xiaomi goods that make your home a voice-prompted, AI-boosted environment. Because of a regional production hub in Shenzhen, these goods, which involve amplifiers, fridges, rice cookers, and hoovers, are too low-priced. China’s production benefit and privacy hesitations of the US enable China to get a 60-40 benefit currently, and the writer forecasts it will raise to 80-20 within five years.
The fourth and last wave is autonomous AI. Until now, we could not even get near to a technology that provides robots human-like mind, and we might not reach it in the future. However, we have drones that are getting better and machines that can understand the color of a good strawberry and softly harvest them. Google and Tesla also work on changing our roads with driverless automobiles which will be unveiled in the future.
Thus, the US has now a great captainship in autonomous AI that the writer gives around 90-10, however, China desires to catch up. The Chinese government is extremely motivated and enterprising in processing AI-friendly policies, thus it will be simpler to apply this technology wider. China is already constructing a highway and a whole city in the size of Chicago specifically made for AI automobiles. Therefore, within five years, it will be near to a 50-50 divide.
Chapter 6 – Specialists consider that AI will cause a utopia or a dystopia.
Nowadays, when economists and scholars discuss how a world with an AI economy would be, they are mainly categorized into two parties.
Renowned geneticist and scholar, Ray Kurzweil, is in the utopia party. He perceives machines as superior means for humans to improve our bodies and minds, enabling us to be more intelligent and live longer. Like him, AI scholar Demis Hassabis perceives AI as a means which would enable people to cure any disease and solve any issues such as climate change at last.
In the dystopia party, stands entrepreneur Elon Musk and physicist Stephen Hawking who consider AI’s potential as a very significant threat to human beings. For instance, an AI code could be wanted to solve climate change and recognize not having human beings in the world as the best choice.
Views vary among economists too, and lots of discussions have rooted in a 2013 study from Oxford University, which discovered that 47 percent of US occupations will be in danger within the following 20 years because of the boosting automation. Lee’s disputed forecast is that over the next 15 years, AI will reach the technical skill to erase 40 percent to 50 percent of occupations in the US.
A lot of companies would be happy to decrease costs and boost profits if they could do it by automating some duties. And this introduces us to a significant distinction between the Oxford study and the reports that occurred after the study: The majority of the automation that AI can do now enables some duties to be automated, however, not all of the occupations.
For instance, an automated tax consultant could do specific duties, such as computing tax returns and controlling for inconsistencies, however, it cannot have detailed conversations with customers.
Taking the distinction between some duties and whole occupations into consideration, more reports came. The Organization for Economic Cooperation and Development stated that just nine percent of American occupations were in danger because of automation. According to a 2017 paper by PriceWaterhouseCoopers (PWC), 38 percent of American occupations were in danger, and McKinsey Global stated that approximately 50 percent of duties in the world are “already automatable”.
This is a fairly wide range, and it is a major reason why economists stay in disagreement on this point. The writer is likely to go with the PWC report but thinks that the real quantity of removed workers may even be more.
This is since the reports did not think about the ground-up removal, which would be observed in businesses such as Smart Finance and Toutiao, which do not have any loan officers or editors. Therefore, these businesses will not be removing workers to place automated solutions, instead, they will remove loan officers and editors by not opening positions for these occupations from the beginning.
Lee ends by telling that he is honored with his achievements as an AI scholar and scientist. However, he feels bad about his priorities being changed. “Instead of seeking to increase the potential of the human brain”, he mourns, “I should have sought to understand the human heart.”. Although some people concern that an AI economy would cause destructive losses in occupations, if we would change our values to favorable human-to-human occupations like caregiving and community-related jobs, we may become an even greater community and make the world a better place. He also talks about some notions like “universal basic income”, “guaranteed minimum income”, and “job sharing”.