China’s AI: Should We Worry?

China’s ruling Communist Party (CCP) is making advanced technology, like artificial intelligence, a cornerstone of its industrial policy. In the fourth industrial revolution the world is now facing, China wants to lead. Some say they are leading mainly on the regulatory side. But what about the rest – such as market share and tech influence in other countries?

Washington is watching. For now, the U.S. is in the lead. Sanctions against companies like Huawei have put up detours, but Beijing may be using those detours to map out new roads to build a massive, soup to nuts, tech ecosystem.

Georgetown University’s Institute for Technology Law & Policy teamed up with the Information Society Project of Yale University on Friday for a symposium to discuss these topics. The web-only event, titled “The Geopolitics of Chinese AI” may have, at first, sounded like praise for the CCP’s recent crackdown on its tech firms – from demanding compliance on basic things like driver’s licenses from ride-sharing company Didi, to putting something akin to a kill switch for mobile games for kids in mobile game apps. But none of the panelists believed a China model was something to imitate in the West.

“Around the world, many teenage parents might like rules on video game limits, and limits on tuition, but these algorithmic rules would be difficult if not impossible to impose in the U.S., for economic reasons, let alone political and social reasons,” said Simon Chesterman, one of the panelists and a tech lawyer teaching in Singapore. “I have a blind taste test. It is really striking that some of the more positive initiatives are coming out of China, but I would like to see what really happens when the rubber hits the road,” he said.

For China’s AI, the thinking is that the CCP has decided that protecting certain values is more important than protecting a market. While China may have signed onto agreements with UNESCO that AI would never be used for unreasonable surveillance or against human rights, in reality, China’s human rights record and use of advanced technology to track and repress religious and ethnic minorities have led to sanctions. The rules they tout publicly and want you to believe are their standards for AI are almost identical to that of the EU on paper, but reality tells a different story.

“This kumbaya language does not serve anybody’s ultimate interest,” said Kendra Schaefer, Partner and Head of Tech Policy Research at Trivium China. “China really believes they do all those things…by their definition and viewpoint of those things.”

What’s China’s Plan?

It would be easy if it was just one issue, but it is a bunch of issues, said Schaefer. “There is a giant list that we keep adding to about the motivations behind all these regulations.”

One way the CCP wants to use artificial intelligence is for managing – or better yet, controlling — social cohesion. They maintain their power by maintaining stability. (Recent and surprising protests in Shenzhen against Covid policies, perhaps organized online, could be shut down with AI, making it nearly impossible to organize online).

Beijing has also had enough of its tech entrepreneurs getting too powerful so quickly. So the thinking is that China wants to now move from a period of barbaric growth to a period of state-managed compliance, or “sustainability.”

Only about half of the drivers in a Didi car are even legally able to drive, for example. So part of the AI push is about managing compliance. It seems positive on one hand for providing assurance to Didi ride-share passengers, but detrimental to a society that is constantly monitored and controlled.

Beijing’s definition of innovation is changing.

“Innovation used to just mean making new stuff and being on the forefront of any kind of tech. Now they are focused more on making things from nothing, rather than just scaling what’s already in existence,” Schaefer said, a very Tao Chinese principle of making something from nothing. “They want the core breakthrough technologies in China. Those core breakthrough technologies are things the U.S. has sanctions on. China wants the core tech companies to help start-ups to break the U.S. chokehold,” she said.

That comment was one of the more interesting in the two-hour symposium. Venture capital firms like Sequoia Capital and Intel Capital are big investors in this sector of the nascent Chinese high-tech market.

Warning: Sanctions Created Detours, not Roadblocks

Entity List bans on U.S. computer hardware companies from doing business with the likes of Huawei, others, have necessitated Beijing’s efforts to build out a homegrown high-tech ecosystem, panelists claimed.

Moreover, Beijing has often copied (or stole) Western innovation and IP until it could sufficiently takeover the industry domestically itself.

“Beijing is supercharging its support of the semiconductor industry. They are making mostly low-quality chips now. But it used to be that only a couple of big firms focused on semiconductors. Now the policy has switched, and Beijing has said ‘everyone needs to make semiconductors’. Tencent is getting into the game. They’re pushing Alibaba to get into the game. Beijing is focused on small to mid-sized enterprises (SME) to make one specific type of chip, like for automotive. They’re pushing money into universities and connecting them to the market more closely. That includes policies for R&D that then link to SMEs that want to get into this stuff and can afford top tier talent, get them ready to go. None of this is guaranteed and we won’t know for a decade if they have solved the problem or not, but if China was not subject to such sanctions, they would not have done this. And when the wheels of China’s economy get grinding down a certain path, they usually get where they are going. It is something worth being concerned about.” – Kendra Schaefer, Partner and Head of Tech Policy Research, Trivium China.

Anupam Chander, professor of law and technology at Georgetown, noted that Huawei used to be a niche player, making lower-end products, and has grown in 10 years to be a global 5G juggernaut. “Huawei is leading-edge,” he said.

Exporting the Surveillance State?

Is China exporting its surveillance state model? This was another question raised by panelists. No one seemed to be all that worried that this was the case, or that if it was, they would get away with it. The message was that if authoritarian countries wanted to create digital prisons for their people, they can buy low-cost China equipment, or they can buy U.S. and European equipment.

“We hear a lot about China’s export of its digital model. We know that China is really good at providing high performance, cheap surveillance technology. But to what extent is China’s government intentionally going around and trying to find other governments who are trying to do the cut and paste of China’s regime? I’m hard pressed to find that. The fact is that there is demand for surveillance tech and the question is how much is demand and how much is being pushed by China. Xi Jinping’s leadership is about making the world safe for autocracy, but that is unique to China.” – Samm Sacks, Senior Fellow, Yale Law School

Chander said it was worth noting that the “great firewall of China” – where it is next to impossible to access the non-China internet —  was built by Cisco Systems. It was yet another example of U.S. tech companies helping build a Frankenstein monster.

China remains around 10 years behind the U.S. in tech, but that gap can close in a hurry, thanks not only to the smarts of Chinese scientists and computer programmers but untold sums of money and advice from Silicon Valley and Wall Street.

“This is a dangerous moment in a way,” Chander said. “So much concern about how each country is racing ahead and worry that one country is becoming the superpower in AI. We want the world to have the strongest technologies and I’m happy that Chinese scientists are producing strong technologies.  Using technologies to target specific groups is bad and we point our finger at China, and rightfully so. We have to make sure the more harmful aspects are properly curbed.”

Who wins the AI race? Nobody knows.

“One thing we know about the AI race is we are bad about technological forecasting. We thought the same thing in the 80s. We said the same about nanotechnology and now nobody’s even talking about it anymore,” said Jeffrey Ding, a fellow at Stanford University’s Center for International Security and Cooperation.

“AI will be a transformative technology. Who will maintain control over the AI? That is one of the challenges when we think about AI governance. Europeans have their civil and human rights-based approach. The U.S. takes a free market approach. China has a state security approach. China has been talking about using AI in the legal system, using one click, predictive judgements. That won’t be used in the West…but that technology should give us pause. We should be cautious assuming (we cannot) extrapolate China’s system into other systems. We will be talking about this topic for years.” – Simon Chesterman, Senior Director, AI Singapore