Intro: China’s AI Dream
The Chinese government has stated its intention to turn China into “the world’s primary AI innovation center” by 2030. If China were to meet that goal, it would have a profound impact on the global economy, geopolitical order, and everyday life for citizens around the world.
But what does it take to strengthen national prowess in AI? What are the building blocks or stumbling blocks for China to realize its AI ambitions?
In this series of digital interactives, we will take you through the main building blocks for creating and deploying AI, illuminating unique aspects of the Chinese ecosystem for each one. Before doing that, we need a rudimentary definition and model for AI today.
What is AI today?
AI is often defined as machines able to perform tasks that previously required human intelligence, such as visual perception, language processing, pattern recognition, and complex decision-making. AI can be used to increase accuracy on a given task, improve efficiency in a process, or substitute for human labor.
Much of the AI deployed today is based on a set of recent advances called “deep learning.” In deep learning, AI practitioners construct artificial neural networks: algorithms that are loosely modeled on neural connections in the human brain. They then feed large amounts of digital data into those networks, using that data to train the neural network to perform a specific task with superhuman accuracy.
AI is an “omni-use technology,” one that can be applied to thousands of different tasks: diagnosing a disease, identifying a face, driving a car, conducting a drone strike, predicting the weather, making a home loan, or conducting surveillance. This is why some prominent AI practitioners have called it “the new electricity,” a technology that will permeate almost all aspects of society, enhancing many traditional industries and opening doors to entirely new ones.
Doing that at scale requires bringing together the key building blocks of a functioning AI ecosystem: data, talent, companies, and public policy. The illustrations below provide an analogy for how these pieces fit together.
If AI Were a Rocket…
Data is the rocket fuel. For many AI applications, without it you go nowhere and the more you have, the further you go.
AI talent—machine learning PhDs, algorithm engineers—are the rocket scientists, designing the rockets and adding the fuel.
AI companies are the rocket companies that gather the data, employ the scientists, and commercialize the rockets.
The government plan is the policy environment that rocket companies act within, including subsidies, regulations on launches, building codes, and public partnerships.
China’s AI ecosystem exhibits strengths and weaknesses for each of these building blocks:
- While China’s data environment lags in diversity, it is particularly strong in the depth of data on each Chinese user.
- In terms of talent, Chinese AI labs lag far behind their US counterparts in best-in-class research, but Chinese-educated researchers are active participants in elite US and international labs.
- China’s AI startup environment has seen an enormous boom in recent years, and is now being shaped by alliances and rivalries between the country’s tech giants and lead investors.
- While China’s national AI plan has spurred a boom of local activity in AI applications, it remains to be seen if the plan can generate the research breakthroughs desired.
Click on the buttons below or use the sidebar navigation menu for a deep dive into each of these four key building blocks.
In the coming months, MacroPolo will further expand the ChinAI project. We will examine a fifth key AI building block (semiconductors), as well as analyze the ethical and safety concerns raised by the deployment of AI systems. For more on these future topics, click on “What’s Next” in the sidebar.