I picked up this book since the subject of Artificial Intelligence (AI) is close to my heart and was one of the main reasons, I dived into computer science as a major subject of study way back in 1985. Back then, AI was divided in two camps the “rule-based” / “expert systems” camp and the other the “neural network” camp. Both these competing approaches did not see eye to eye and largely the rule-based camp prevailed till abundance of computing power and data catapulted the neural network approach to where we are today.
Born in Taiwan and educated in the USA, Dr. Kai-Fu Lee, is an eminent AI expert, a thought leader in the field and a venture capitalist (VC) heading Sinovation Ventures. Exposures with Apple, Silicon Graphics, Microsoft (created Microsoft Research in China) and then President Google China he is ideally placed to evaluate the two tech ecosystems in Silicon Valley and China.
There is no dearth of books on AI, but AI Superpowers: China, Silicon Valley and the New World Order, stands out since the book has been written based on extensive on ground experience of the author in the Chinese tech-ecosystem. Hence the book’s China perspective prevails throughout. At 232 pages the book is a fast and an easy read even for non-tech readers.
AI, is one of the key technologies affecting us in our everyday lives and there is no doubt that the best AI minds of the world are in US and China. The book can be divided in two major parts, 1) The Rise of China as an AI superpower, the AI ecosystem in both US and China and how tech giants in both countries are competing for technical and economic advantage 2) Economic impact of AI. The first part of the book is extremely informative and highly readable. I could disagree with many issues raised in the second part of the book where he talks about the economic impact of AI. His battle with cancer may have shaped many of those views on economic impact of AI. Though he derives his views on economic impact from various studies but, it would have been better if the views of professional economist were obtained to support his arguments.
The book delves into Deep Learning, a branch of AI where given enough data and compute, the algorithms learn themselves and how this technology is being used in present day China. He correctly appreciates the usefulness of the technology and simultaneously debunks the myth around “Super-AI” and rightly says that at present it is science fiction and is a long way off, if at all feasible.
Lee, in his book discusses in detail as to how the four critical ingredients i.e. “abundant data”, “hungry entrepreneurs”, “AI scientists”, and an “AI-friendly policy environment” is helping the Chinese side to tilt the balance of power in AI Supremacy. He explains each of the four ingredients and compares them across US and China ecosystems. Abundant data, he says, is of very high quality given the ubiquitous Chinese super app “WeChat”. A key advantage to China in collection of large volume of data is that “companies are less constrained by government privacy rules” and “government itself helps with data collection” compared to West where stricter data privacy laws exist. Lee says “there’s no data like more data”.
As per Lee, “China’s start-up culture is the yin to Silicon Valley’s yang: instead of being mission-driven, Chinese companies are first and foremost market-driven”. Their goal is “to make money” and to achieve that “they create any product and adopt any model” to the extent of even “copying” successful business models across the world making Chinese entrepreneurs more aggressive that their western counterparts. I somehow disagree with his justification of the “copycat” philosophy, which in my view is plain and simple state sponsored intellectual property theft. Lee spells out how the Chinese government focus on AI has energised the provincial and local governments to spend on the support infrastructure to provide funding, incubators etc and compares it to the hands-off, market-based approach of the US.
From a learning standpoint for non-tech readers, the description of the “Four Waves” in which AI revolution will manifest itself is illuminating. These are First Wave – Internet AI, Second Wave – Business AI, Third Wave – Perception AI and the Fourth Wave – Autonomous AI wherein each wave uses the power of AI in a different manner. He also talks about “O2O” (Online-to-Offline) AI applications where online actions result in offline services and blended environments of OMO (Online merge Offline).
Coming to the economic impact of AI, Lee makes various points on job losses, loss of self-worth and rise of global inequality concentrating power in the hands of US and China. As far as job losses due to AI adoption is concerned, Lee peruses four studies which assessed the loss of jobs. First, 2013 study by two Oxford University researchers Carl Benedikt Frey and Michael A Osborne who predict automation of 47% US jobs in the next decade or two. Second, the 2016 research by Organization for Economic Cooperation and Development (OECD) which contrasted the Oxford prediction and said that just 9% jobs were at high risk of automation. This substantive difference was on account of “occupation-based” approach by Oxford and “task-based” approach by OECD. The OECD team broke down each occupation into its many component activities and looking at how many of those could be automated. Third, the 2017, researchers at PwC used the task-based approach to produce their own estimate, finding instead that 38% of jobs in the United States were at high risk of automation by the early 2030s. Fourth, the McKinsey Global Institute study which was somewhere in the middle. As per them if there is rapid adoption of automation techniques 30% of work activities around the world could be automated by 2030, but only 14 percent of workers would need to change occupations. Lee discusses various reasons for the substantial difference between the studies and finally agrees with the PwC prediction. In a decade’s timeframe other studies available in public domain peg the figure between 10-25% job replacement.
He also observes that virtually all the gains will go to a small number of wealthy tech titans and competitive markets will be overturned and we will “see greater and greater concentration of astronomical sums in the hands of the few.” This is a debatable conclusion and I am sure professional economists would be having their views on this subject.
Many of the proposed technical solutions for AI-induced job losses, as per Lee, fall into three categories – the 3Rs – “Retraining workers”, “Reducing work hours”, or “Redistributing income”. He discusses each one of them in detail including the concept of Universal Basic Income.
Finally, the key message in the book is that in the $15.7 trillion global AI ecosystem, China is poised to be an AI superpower due to a focussed government, strong entrepreneurial outlook, high quality, large volume of readily available data and highly skilled AI experts. There are concerns regarding job losses due to AI adoption, but there are solutions to overcome them. He feels human-to-human jobs like caregiving and community-based work etc will emerge and create a better society.
Overall the book comprises Lee’s insights about the US-China interplay of AI related technologies and paints an extremely rosy picture of China’s strengths. In my personal view, such conclusions maybe premature, but one thing that is clear is that China is a serious player in the game and must be taken seriously.