Artificial intelligence is often presented as a story of technological innovation, but its development is also a story of power, economics, competition, and human transformation. In AI Superpowers: China, Silicon Valley, and the New World Order, Kai-Fu Lee examines how artificial intelligence is reshaping the global economy and argues that China and the United States are positioned to become the two dominant forces in the AI revolution.
Book cover image: amazon.com
Published in 2018, the book remains an influential introduction to the geopolitical dimensions of artificial intelligence. Drawing on his extensive experience in AI research and the technology industry, including leadership roles at Microsoft Research Asia and Google China, Lee offers an insider’s perspective on the contrasting strengths of the Chinese and American technology ecosystems. His central argument is that the era of AI has moved beyond simply inventing algorithms: data, computing power, entrepreneurial energy, and the ability to deploy technology at scale have become critical sources of AI power.
China vs. Silicon Valley
One of the most compelling aspects of AI Superpowers is Lee’s comparison of China and Silicon Valley.
For much of the internet’s early development, Silicon Valley was regarded as the unquestioned centre of technological innovation. Chinese technology companies were sometimes dismissed as imitators of American products. Lee challenges this perception, arguing that China’s technology ecosystem has developed its own strengths.
The United States, he argues, has traditionally possessed an advantage in fundamental research, world-class universities and elite AI talent. China, however, has advantages in areas such as data, entrepreneurial competition, government support and rapid implementation.
This distinction is central to Lee’s argument. The AI race is not simply about developing the most sophisticated algorithm. It is also about having the data, infrastructure, capital and users necessary to deploy that technology in the real world.
Lee therefore describes China as particularly well positioned for what he calls the “age of implementation.” The ability to rapidly test, improve and deploy AI applications could become as important as making fundamental breakthroughs.
The Four Waves of AI
Perhaps the most useful framework in the book is Lee’s model of the four waves of AI: Internet AI, Business AI, Perception AI and Autonomous AI.
1. Internet AI
The first wave involves AI systems that learn from people’s online behaviour. Recommendation engines, targeted advertising and personalised content are examples of this form of AI.
Most people interact with Internet AI every day, often without consciously recognising it. What appears on a social-media feed, which products are recommended to us and which videos appear next are all influenced by algorithms analysing user behaviour.
2. Business AI
The second wave applies AI to structured business data. Banks, insurance companies, hospitals and other organisations can use AI to identify patterns and support decision-making.
This wave has important implications for professional employment because many administrative and analytical tasks can potentially be automated or augmented by AI.
3. Perception AI
Perception AI involves machines interpreting the physical world through technologies such as computer vision, speech recognition, and sensors.
This moves AI beyond the digital environment. Machines can increasingly recognise faces, understand speech, interpret images and interact with their surroundings.
4. Autonomous AI
The final wave involves machines capable of independently interacting with and navigating the physical world. Autonomous vehicles, robots and drones are examples.
This wave has potentially profound consequences because it could affect both white-collar and physical occupations.
Lee’s four-wave framework remains one of the book’s most useful contributions because it gives readers a simple way of understanding how AI capabilities can move from online environments into businesses and eventually into the physical world.
The Importance of Data
A major theme throughout the book is the importance of data.
Lee argues that modern AI systems can become increasingly powerful when they have access to enormous quantities of high-quality data. China’s enormous digital economy therefore becomes an important competitive advantage in his analysis.
However, the relationship between data and AI is more complicated than simply having “more data.” Data quality, computing infrastructure, algorithms, technical talent, privacy regulations and the ability to turn data into useful applications all matter.
Nevertheless, Lee’s emphasis on data was particularly prescient. The development of today’s generative AI systems has reinforced the importance of large datasets, computing infrastructure and access to enormous amounts of computational power.
AI and the Future of Work
The most important—and perhaps most unsettling—part of the book concerns employment.
Lee argues that AI could eliminate or substantially transform millions of jobs. Unlike previous technological revolutions, which often automated physical labour, AI can increasingly perform cognitive and knowledge-based tasks.
This creates an uncomfortable possibility: what happens when machines become better than humans at many of the skills for which people have traditionally been paid?
Lee does not believe that simply retraining everyone for technical jobs will solve the problem. Instead, he encourages society to rethink the meaning of work and place greater value on qualities that machines struggle to replicate, including empathy, compassion and human connection.
This is particularly relevant to education. If AI can increasingly provide information, generate text, analyse data and perform routine intellectual tasks, education must focus not only on knowledge acquisition but also on critical thinking, creativity, collaboration, ethical judgement and responsible use of technology.
The Human Cost of Technological Progress
One of the book’s most powerful ideas is that the benefits of AI will not necessarily be distributed equally.
AI could generate enormous wealth and increase productivity while simultaneously creating economic insecurity for people whose work becomes automated. This creates what Lee describes as a major social challenge: society needs to determine how the wealth created by AI can benefit people whose jobs and livelihoods are disrupted.
This makes AI Superpowers more than a book about technology. It is ultimately a book about social policy and human dignity.
Lee considers solutions including new forms of social support and a greater emphasis on activities that strengthen human relationships. His argument is that societies should not measure human worth exclusively through economic productivity.
The Wisdom of Cancer
The book takes an unexpected personal turn when Lee discusses his diagnosis with advanced cancer.
This section changes the tone of the book. After spending much of the earlier chapters examining technological competition, economic disruption and the future of AI, Lee reflects on mortality, relationships and what really matters in life.
His experience reinforces an important distinction between intelligence and humanity. A machine may be able to calculate faster, recognise patterns more accurately or process more information than a human being, but human beings experience relationships, compassion, vulnerability and love in ways that cannot simply be reduced to computational performance.
This personal dimension makes the book considerably more memorable.
Strengths of the Book
One of the book’s greatest strengths is its accessibility. Lee manages to discuss complex AI concepts without requiring readers to have a technical background.
His professional experience also gives the book an unusual perspective. He understands both the technological foundations of AI and the commercial realities of deploying technology.
The comparison between China and the United States is another major strength. Rather than assuming that Silicon Valley will automatically remain dominant, Lee asks readers to examine the structural advantages that could allow China to become a major AI power.
The book is also commendably broad. It connects AI with economics, employment, education, geopolitics and human values rather than treating artificial intelligence simply as a technological development.
Limitations of the Book
The biggest limitation is its publication date. AI Superpowers appeared in 2018, before the emergence of today’s mainstream generative-AI systems and the extraordinary public adoption of tools such as ChatGPT.
Consequently, some of Lee’s timelines and predictions should be read as forecasts rather than established facts. The AI landscape has changed substantially since the book was written.
The book’s strong focus on China and the United States can also make the global AI landscape appear more binary than it really is. Europe, the United Kingdom, Canada, India, Japan, South Korea and many other countries have important roles in AI research, regulation, infrastructure and adoption.
There is also a danger in describing AI development primarily as a competition between superpowers. Competition can stimulate innovation, but AI also presents problems—such as safety, misinformation, privacy, employment and governance—that require international cooperation.
These limitations do not make the book obsolete. Instead, they provide an interesting opportunity to compare Lee’s 2018 predictions with the AI developments that have occurred since its publication.
Why AI Superpowers Still Matters
Perhaps the greatest achievement of AI Superpowers is that it encourages readers to think about AI at a systemic level.
Instead of asking only, “What can this AI tool do?”, Lee encourages us to ask:
- Who controls the data?
- Who develops the technology?
- Who benefits economically?
- Which jobs will change?
- How should education respond?
- How should governments regulate AI?
- What happens when technological progress increases inequality?
- What human qualities become more important as machines become more capable?
These questions are even more relevant in the current AI era.
Today’s AI conversation often focuses on individual tools and their practical applications. Lee’s book encourages us to step back and consider the much larger transformation taking place.
AI Superpowers: China, Silicon Valley, and the New World Order is an insightful and accessible introduction to the global politics and economics of artificial intelligence.
Its central comparison between China and the United States provides a useful framework for understanding why AI leadership depends on much more than scientific breakthroughs. Data, talent, entrepreneurship, investment, infrastructure and government policy all influence a country’s ability to develop and deploy AI.
More importantly, Lee reminds us that the AI revolution is ultimately about people. The greatest challenge may not be creating intelligent machines but creating a society capable of managing the enormous economic and social changes that intelligent machines bring.
Although some aspects of the book have inevitably been overtaken by developments since 2018, its central questions remain remarkably relevant. AI has moved from a specialist technological field into everyday life, education, business and public policy. The challenge Lee identified—how humanity can benefit from AI without losing sight of human dignity—has become even more important.