About the author
Kai Fu Lee is born in Taiwan from Chinese parents, he emigrated to the USA at the age of 11, with a passion for mathematics. He did a PhD at Carnegie Mellon University, “the incubator of cutting-edge AI research” and became passionate about machine learning and voice recognition. After having held executive positions at Apple and Microsoft, he went to China and became the first Chinese investor in venture capital with his company Sinovation Ventures. He advised many students and trained more than 5000 specialists in AI, including several senior executives of Baidu, Alibaba and Tencent.
His 30 years of experience makes him one of the world’s leading experts in AI. His book traces the evolution of AI from 1970 to 2017, when the Chinese government declares AI a top priority. “Sillicon Valley and China have an ecosystem born of very different cultural backgrounds” and this play a significant role in the approach and evolution of each country. China is known to have copied in many areas, often in low-end items, but these copiers have learned, assimilate with impressive speed and have launched into the creation of products and services tailored to the needs of the Chinese market. Indeed, China has been able to collect masses of valuable data that feed the algorithms in their purchases, their habits and their movements, ensuring a global advance compared to the rest of the world. The deep learning feeding on all these data to recognize the images, understand the spoken language, propose solutions after analyzing and cross-checking databases.
To date, China has an undeniable lead in the collection of these data with Tencent (owner of WeChat), as well as Alibaba with e-commerce and strong of more than one billion users with the intensive use of mobile (mobile only). But AI will affect all industrial structures, all areas of life and its progress will harm the professional environment, massive unemployment, greater inequality and bringing social crises. “AI would replace 40 to 50% of US jobs within 15 years while fortunes would accumulate at the big AI companies.”
The author makes a fine analysis and gives advice on the measures to be taken by the governments in all countries and asks them to think now about the sectors that will be affected.
In 2016, 280 million Chinese viewers follow an Alpha Go confrontation with the best Korean player who succumbs to the machine. The Go is a 2500 years old game in China and still practiced in the 21th century. In 2017, the best player in the world, a young Chinese of 19 years old (Ke Jie) was beaten by the machine. With stupefaction and awareness, China governs immediately by mobilizing all its actors to meet the challenge of AI. The author compares this moment to the one experienced by the US when the first Soviet satellite “Sputnik” was launched in 1957. “This triggered the creation of NASA, the teaching of mathematics subsidized on a large scale and allowed 12 years later to send the first man to set foot on the moon: Niel Armstrong! ”
It is thanks to deep learning and machine learning, that the capabilities of AI have allowed a machine to win the Alpha Go. The author specifies the need to have “a large mass of data, experts, engineers, calculation capabilities out of the ordinary, colossal sums to advance everything”. It is indeed in this race that all the means have been put in China, as a proof the current level of AI in China. After Alpha Go, China launches a major plan to develop AI. Objectives are set for 2020 to 2025, the stated goal is to be the world leader by 2030.
The 4 waves of IA
Each wave will spread out in time, each one will exploit in its own way the power of this technology to submerge us. The successive waves will each bring facilities in all fields, each of them feeding on the previous one, affecting all economic sectors and our daily life. Already online AI and professional AI are present; the financial field is an example with the algorithms at work on the stock markets.
1st wave : AI online for more than a decade
Around 2012, the internet has propelled us to it: our video choices, purchases, our questions in all areas have been recorded, this data allows to display the target of the company that will give a positive image to customers, and will know our habits, our preferences, while the customer will get among other things personalized offers, “the algorithms are transformed into a tool for recommendations. “This wave remains confined to the technological sphere and the digital world”, Chinese and US are equal according to the author, who sees China in the lead in 5 years.
2nd wave : professional AI
According to Kai Fu Lee, this is the only field where the US domination is unquestionable. All the labeled data accumulated over the years in the insurance, banking and hospital sectors will allow algorithms to make particularly effective optimizations.
“As early as 2004, companies (Palantir – IBM) were advising companies and governments on big data”. The author sees a great evolution in China for the medical field, the algorithms would establish the dignostic which in the country would decrease the inequalities between big cities well equipped and countryside under equipped in this sector; similarly in the judicial sector to guide the magistrates, to pronounce the verdict and “to contribute to put in the right way the customary judges. At this stage, the US is in the lead (90/10) thanks to optimizations in banking and insurance among others. China will be in the lead within 5 years (70/30) by taking the lead especially in services.
3rd wave : perceptive AI
The digitization of the physical world will be implemented in the commercial sectors as a priority: it is necessary to rethink all human activity, to recognize objects, voice and sound data. The author warns that algorithms “will perform the bulk of interactions with the customer” and gives Exemples: Amazon Echo that digitizes the sound environment of the home, the city brain of Alibaba that digitizes the flow of traffic or Apple and Face ID cameras that do the same with the faces. In 2017, during the 1st visit of Trump in China, his speech in English on large screen began to speak Chinese. The AI is transforming the world. The Chinese company Yflytech had specialized in the field of AI to train its algorithms, respecting all the characteristics of the voice: intonation, pronunciation, all recalibrated in Mandarin. In the next few years, cities and supermarkets will be transformed, the author already cites KFC, associated with Alipay, which has installed in some of its restaurants a system that allows facial recognition, with a terminal recording the order, scanning the face, ensuring the presence of the person in the premises and debiting his account.
Moreover, this process will be used to educate, to detect delays or talents, to follow each one and to take in hand the delays of the pupils. The AI will be in the homes, the teachers will ensure the lectures, or will make individual sessions and many other fields will benefit from this process. That implies to recover important data of the physical world. “the US is reluctant, china is more open and has passed a cyber security law. Europe having adopted a more restrictive approach with the RGDP.” From this perceptive, AI needs all kinds of devices with sensors. In anticipation, China has created “the made in Shenzen, by rallying in one place of the chain studies, researchers and manufacturers of intelligent devices; with the largest network: 85 Billion connected objects, various products, reduced prices, amassing important data and preparing the 4th wave of autonomous AI.
4th wave : autonomous AI
For the author, the impact will be more important, it is a culmination: the machines were automated and they will be autonomous. To quote the example: harvesting strawberries in California by a machine guided by an algorithm picking without damage the ripe strawberries, the autonomous robots that prepare the orders at Amazon and eventually drones that will be able to supply isolated populations or put out forest fires.
On their side, Google and Telsa are working on autonomous cars, as well as Baidu, Uber, Didi and car manufacturers. Our daily life will be impacted: urban spaces, roads and already Chinese territories are modifying traffic lanes, adapting supply circuits. In the US, it is the new vehicles that will adapt to the existing roads. Indeed, he US have the most experienced specialists and have a 2/3 year lead over China. The observation made by the author at this stage comparing China/US is as follows: AI : online 50/50 ; pro 10/90 ; perceptive 60/40 ; autonomous 10/90 and the forecast 5 years ahead China/US: online 60/40 ; pro 30/70 ; perceptive 80/20 ; autonomous 50/50. China and US are moving forward, but in the long run developing countries will use them. US are trying to sell their standardized products while China chooses to finance native startups: Tencent or Alibaba in India and South East Asia or even Didi developing in US, Singapore, India or Middle East.
How to react
At the heart of this revolution, the new norm would become full employment for intelligent machines and eternal stagnation for the average worker because the faster, larger AI will surpass humans in physical and cognitive tasks. Solutions are studied taking into account the extent of the destruction of jobs, primarily among the most qualified and among workers, the author describes and gives examples: jobs concerning intellectual tasks in danger: technical translator, scientists etc. Those easy for AI alone : radiologists, accountants, trade unionists or those mixing AI and human as criminal lawyer, doctor and teacher. Similarly for jobs with physical tasks, would be preserved: home help, physiotherapist, hairdresser, aeronautical mechanic and in danger the easy jobs for AI : agricultural worker and truck driver for example.
Different approaches are evoked by the US or by the silicon valley: requalify thanks to continuous training to adapt and bring the necessary competences to the workers, but this would be permanent because of the fast evolution of AI and very quickly arriving at saturation, or redistribute the jobs by decreasing the working time to 3 days per week or share a job which would be held by two people, it would be or partial unemployment or lower wages. Other concepts are debated: the universal basic income which is far from being unanimous, or the minimum income guarantee which would be reserved for the poorest accompanied by heavy taxes imposed on AI companies.
Kai Fu Lee proposes rather a “social investment allowance, with guarantee of care, services, education” that the state would pay to associations “dedicated, creative, volunteers” all this of course after the completion of AI and takes appropriate measures before this time by paying “people caring for the elderly or parents of young children” for example and encourages the private sector to be “pioneer in the creation of human jobs that will be at the heart of this revolution”. The author also suggests to be inspired by other countries, to emphasize the craft knowledge, the culture of volunteering! “With AI, there can only be one winner. US and China will compete to make the most of this technology. Governments around the world must connect permanently to compare US/CHINA/EUR regulation policies and keep an open mind about the different ways to consider AI governance.
After this evolutionary assessment of AI and its disastrous consequences in the professional world (progress on the one hand, destruction on the other, questioning the life of everyone, our environment and our organizations), the author ends on a more optimistic note thanks to his personal experience and the questioning of his own life, due to the occurrence of cancer that has allowed him to become aware of a life other than the one he led. The presence and support of his family during his illness, the strength of love “which we know gives meaning to our lives” and the meeting with a Buddhist monk in a monastery in Taiwan, which led him with wisdom to rearrange the priorities of his life.
As for the end of Kai Fu Lee’s convalescence, he gives us the content “I considered my recovery from two angles, one technical, the other emotional” the two pillars of our future AI.