Inside the US-China War for AI Supremacy

If you want to understand how AI is reshaping global job markets, there may be no better person to talk with than Dr. Kai-Fu Lee.

I met Dr. Lee last month at an event in Malibu. He is chief executive of Sinovation Ventures, was the president of Google China and has been at the forefront of AI development for 30 years.

Dr. Lee’s new book “AI Superpowers” examines China’s super-fast ascent in tech and AI. In Dr. Lee’s view, China has already surpassed the US in some areas of AI implementation.

China today has more AI unicorns (privately held companies with valuations of over $1 billion) than the United States.

What can we learn from China’s success?

China startups operate differently from Silicon Valley

As Dr. Lee acknowledges, many Chinese companies began as copycats of US businesses, shamelessly ripping off American-designed products and business models.

Xiaomi just released AirDots, wireless earbuds that look uncannily similar to Apple’s AirPods (and are five times cheaper).

Chinese entrepreneurs have learned a lesson from this experience, says Dr. Lee: The only way to win is to make your company “uncopyable”.

China’s startups are doing this by 1) Tackling the hardest problems, 2) in businesses that are super capital-intensive, and 3) making it impossible operationally to replicate.

“The challenges and stakes are extremely high in these types of businesses.
 
But if you’ve done it, you’ve won.”

Chinese companies aren’t afraid to rip off what works. Once a business model is proven out, e.g. ride sharing, China’s start-up world tries every possible iteration, competing ferociously to dominate a market: “In China there’d be no chance of having 4 different Yelps for restaurant reviews — one of them would kill the other 3.”
 
Dr. Lee highlights Meituan, one of China’s largest food delivery apps. The app “has changed the way people eat.”

Meituan’s food delivery service covers 300+ cities in China. You order on the app, and meals are cooked for you and ready in 30min, delivered to your door by electrical moped — for an average price of 70 cents.

Building the infrastructure to make a business like this takes huge sums of cash (the company recently raised $3 billion) and is incredibly complex (hiring and training tens of thousands of food delivery employees). But when you reach the scale where you can deliver the best food at the lowest prices, you win. Meituan currently has more than 40% of China’s food delivery market.

“Leave work, click, go home, the food’s at your door. You too will use this if the service gets good enough.”

China has an edge in data

In Lee’s view, AI is shifting from an “Age of Discovery,” where computer science researchers came up with breakthrough insights, to an “Age of Implementation,” in which the key differentiator is access to vast quantities of data.

China has a huge advantage over the US in collecting personal data. It has a bigger population than the US. More mobile transactions. Looser privacy laws. These translate into a massive edge in data to power machine-learning algorithms.

Consider payments. Hardly anyone in China uses a credit card these days. China has over 800 million internet users, vs. 287 million in the US.

China’s data advantage lies not just in breadth (the number of users) but also in the depth of data on each user — the real-world activities of Chinese people that are captured in a digital format.

The most popular app in China, WeChat, sees more than 1 billion monthly active users (MAU). The majority of them pay each other in purely mobile, no 2% transaction fee from credit cards, all commission free.

The enormous amount of data is “rocket fuel” for AI, says Dr. Lee. It provides a multidimensional picture of each Chinese user that gives AI companies the ability to better tailor their services. It gives them an advantage in developing the best tools for machine vision and facial recognition.

China’s government actively supports public infrastructure for AI

The government of China has made AI a top national priority, Dr. Lee explains. They encourage local officials to set up systems conducive to AI development. There’s a push for hospitals to share patient data in an effort to improve AI-powered diagnosis and enable personalized medicine.

Contrast this with the US, where private EHR companies like Epic make it difficult for healthcare systems to share patient data. This limits our ability to use AI to improve diagnosis and realize the potential of personalized medicine.

Privacy concerns in the US place a fundamental constraint on our ability to coordinate data collection. Yet China’s success raises a question: What role should the US government play in our nation’s AI future?

China’s tech economy is not an immediate threat to the US, says Dr. Lee. The US and China run in parallel universes when it comes to the internet economy. Chinese people shop on Alibaba and catch rides on Didi, while Americans use Amazon and Uber. “It’s not a zero sum game.”

Yet Dr. Lee makes a persuasive case that China has already surpassed the US in face recognition, speech recognition, and drone guidance.

These new powers can lead to some creepy “Black Mirror” scenarios. Local governments in China are using facial recognition to track their citizens’ movements.

China’s new social credit scoring system can restrict travel or ban your kids from the best schools.

My takeaway from Dr. Lee’s talk: Rather than fear the rise of China and AI, let’s try to learn from it.

China is a Petri dish for the most ambitious, radical experiments in AI. Advances in facial recognition and self-driving cars will dramatically transform society — in China, the US, everywhere.

As the internet blurs national boundaries, there’s value in studying other economies. China’s booming tech industry provides a glimpse of what’s to come as we hurtle inevitably toward an AI-powered future.