AI Superpowers is a great book for anyone to get an understanding of the modern AI landscape. It describes how this amazing technology can make sweeping changes across the globe, touching many industries and generating extraordinary profits to those who can successfully implement it. Lee provides great depth and analysis in particular of how the battle for AI supremacy is likely to unfold and how the automation of jobs will change the fabric of society.
Kai Fu Lee, is a well renowned engineer, researcher and investor. Having worked in Microsoft, Apple, Google China, and now heading his own VC company Sinovation in Beijing, Lee shares his observations and analysis gained from his experiences.
The first three chapters provide set the context of China’s recent technological developments. Beginning with the spark that ignited them into action in the (chapter 1), the moment AlphaGo defeated the world’s best Go player in 2016. Fearful that this was a sign that China could lag behind USA in a technological race, from this moment China began a concerted effort to pursue deep learning research and implementation. Having been close to the action, he then provides insights into the leading AI entrepreneurs in China, and their style of operation (chapter 2 and 3), including Jack Ma (Alibaba), Liu Qingfeng (iFlyTek) and others.
In chapter 4 he contrasts the similarities and differences between the two world leaders in AI, US and China. The main similarity being that both have managed to create a concentrated ecosystem of talent and capital. While notable differences include:
Lee states that since the break-through research paper published by x in 2006, the AI landscape has entered an age of implementation, in which case China with it’s brute force approach can catch and if not overtake the US. The monopolistic characteristics of AI companies allow leaders in the field to become somewhat unstoppable superpowers. With data being the key ingredient for many AI solutions, and given the fact that with enough data even simple machine learning algorithms can outperform state of the art neural networks, China has a prominent advantage across several domains. The learning nature of AI algorithms is such that the more users a company can attract, the better their products can become and be able to attract further users. As a result these companies that lead the way will accumulate incredulous amounts of money.
The race for AI supremacy across US and China is analyzed across 4 types of AI (chapter 5): Internet AI, business AI, Perception AI, and Autonomous AI. With China and US having their own dominating advantages in each. Each wave of AI threatens to take over jobs that humans undertake in varying degrees. While some fields are in imminent threat such as radiologists and customer support others will be safe such as social workers and CEOs. He bases the threat of AI automation to a job on the level of compassion, and creativity/strategy required. Theoretically he concludes 40-50% of current jobs in the USA may be automated in 15 years. Although the familiar counter argument is that new jobs will be created, he fears unlike the industrial revolutions of the past, this time things would be different and there maybe no new jobs to be retrained into or at least at the mass scale and industry widedness of the changes he predicts.
After dedicating significant time and efforts in AI, Lee shares in chapter 7 a heartfelt story. He describes how he himself became algorithmic like in his daily life. He describes how he was looking at life as one great optimization task. His objective being to optimise impact in his chosen area, and to bring the benefits of his products to as many people as possible, and no doubt making super profits along the way. However, being diagnosed with cancer in 2013, brings profound changes to Lee’s life, and marks a turning point. After receiving tremendous love and care, he realises the great defining factor of humans, and what makes us soo special is our ability to love. Serving and helping another human being brings about an unrivaled sense of self-fulfillment. These enlightenments force Lee to changes his personal mission statement, and fosters within him thoughts that maybe alternatives do exist to the Dystopian jobless futures that seem inevitable.
In the final chapters, Lee challenges, current day AI experts innovator and leaders, to start thinking about how we can harness the goodness of people, and create a symbiosis between man and machine to make miracles happen. In order to achieve this, the current measurements of profits and self-worth derived from work need to be changed. Now, in 2020 it is still hard to see whether these changes will ever take place. For Lee, who has already left his dent in the universe, we hope him and others in his place can help deliver the innovative ideas to move forward into a positive AI future.