This week on “Intelligence Matters,” host Michael Morell speaks with Wall Street Journal reporters Liza Lin and Josh Chin about their new book “Surveillance State: Inside China’s Quest to Launch a New Era of Social Control.” Their new reporting examines how China’s data collection goes beyond that of other countries as China seeks to create a model to export as an alternative to democratic governance. Lin and Chin detail Silicon Valley’s involvement in the buildup of China’s surveillance state and how the technology is used to surveil Uyghurs in Xinjiang.
HIGHLIGHTS:
- CHINA’S ALTERNATIVE TO DEMOCRACY: LIZA LIN: China is probably the only country out there that hopes to use surveillance to create this techno-utopian state. And as we mentioned in our research and in the book, China has this ambition to use the data collected to analyze any future threats to its governance and to identify these threats quickly and do something about it, to create an alternative model to what democracy could offer.
- HOW CHINESE SURVEILLANCE DIFFERS FROM OTHER COUNTRIES: LIZA LIN: “And I think the difference between China and many other countries is that all these cameras are largely state owned. Unlike in the U.S., where you have a ton of Amazon ring cameras that are privately owned, a lot of the surveillance cameras that you see on the street in China are owned by government agencies and largely by the Chinese police. And beyond access to those 400 million cameras, the Chinese government still has access to about a billion smartphones that the Chinese citizens use. And that’s because there are a series of national security and intelligence laws that were put in place in China over the last decade that actually allow the Chinese government to have access to a lot of the information that Chinese tech companies collect.”
- SILICON VALLEY’S CONTRIBUTIONS TO THE CREATION OF CHINA’S SURVEILLANCE STATE: JOSH CHIN: If you think about Silicon Valley involvement, it’s huge. And not just in components. If you think about the financial aspect of things, a lot of U.S. private equity companies and venture capital companies were the first companies to give Chinese surveillance startups a leg up in the game by funding them. So there are a lot of ways that the West and in particular the U.S. has contributed to the development of China’s surveillance apparatus.
Download, rate and subscribe here: iTunes, Spotify and Stitcher.
INTELLIGENCE MATTERS WITH LIZA LIN AND JOSH CHIN
PRODUCER: PAULINA SMOLINSKI
MICHAEL MORELL: Liza, Josh, it is great to have you on intelligence matters. Welcome.
JOSH CHIN: Thank you. It’s a real pleasure to be here.
LIZA LIN: Thank you for having us.
MICHAEL MORELL:You guys are the authors of a terrific book. It’s called Surveillance State: Inside China’s Quest to Launch a New Era of Social Control. Congratulations on the book. I read it from cover to cover. I found it interesting. I found it insightful. I found it readable and more than a bit scary. I’d like to start by asking you guys how you define a surveillance state. How do you think about it in terms of a definition?
JOSH CHIN: I think the way we thought about a surveillance state is a state in which governance, the act of managing society is driven by and really indivisible from advanced surveillance technologies. You see this really clearly in China, where the Communist Party there has melded its traditional approach to governing and controlling China with cutting edge technologies like facial recognition, techniques like data mining, in order to create a kind of more nimble form of authoritarianism.
MICHAEL MORELL: It’s interesting that you talk about advanced surveillance technologies, because one of the questions I wanted to ask is, how does surveillance in China today compare to old Mao style surveillance where neighbor was spying on neighbor? How do you about think about that?
JOSH CHIN: It’s interesting, when we started looking into this, as you know, one of the things we discovered was that really I mean, just to a certain degree, all states are surveillance states going back to the beginning of states. In the sense that anyone who’s trying to govern a large number of people needs to have information about those people in order to govern them. James C. Scott sort of famously described this as rendering societies legible. And so this has been all the way back to the Romans conducting surveys so they knew who to tax and who to draft for the Roman Legion. It’s as old as time, basically.
I think in the 20th century, what you started to see with Mao and with other totalitarian leaders was this utopian idea. This idea that you could collect enough information to mold society in a scientifically ideal way. I think the difference between Mao and what Mao was aiming for and the types of information he was collecting and what you see now is mostly a difference in volume and value of data collected. Mao obviously was limited to human data collection. So he had a lot of human spies. He had this vast domestic spying network, hundreds of thousands of people, and of course, was encouraging other people, regular people to inform on each other. And the Communist Party now, I mean, they still do some of that, but they have these other just vastly more powerful tools that are so much more efficient at collecting and analyzing information. And so they just have the ability to consider this utopian project of engineering society in a way that Mao never actually really could.
LIZA LIN: I just to add on to that, the intent for China to surveil its citizens was always there. But what we really saw in the last kind of 10 to 15 years was the ability because there’s been huge advances in technology. For example, the advent of deep learning and the realization that you had these fast processing chips that could crunch that data in a very efficient manner. So crunched down large amounts of data and draw analysis from that data as per what Josh mentioned. That really actually spurred the development of China’s modern day surveillance state.
MICHAEL MORELL: What are the key components of the surveillance system in China? How do you think about that?
LIZA LIN: To me the way China’s surveillance state stands out is in two areas. One is the breadth, the data collection. And when we talk about the breadth of the data collection, I refer you to the amount of touch points where the Chinese government can actually extract data from its citizens. So the Chinese government has access to probably about 400 million cameras across the country.
And I think the difference between China and many other countries is that all these cameras are largely state owned. Unlike in the U.S., where you have a ton of Amazon ring cameras that are privately owned, a lot of the surveillance cameras that you see on the street in China are owned by government agencies and largely by the Chinese police. And beyond access to those 400 million cameras, the Chinese government still has access to about a billion smartphones that the Chinese citizens use. And that’s because there are a series of national security and intelligence laws that were put in place in China over the last decade that actually allow the Chinese government to have access to a lot of the information that Chinese tech companies collect. And the difference in China versus other countries is that a lot of citizens are largely reliant on Chinese apps. And in particular, this one app called WeChat, which is made by and run by a company called Tencent. And on this app WeChat, it kind of functions quite similarly to a mesh of Facebook, Instagram, Google and Amazon. And on top of that, they are mobile payments. So the Chinese government, through this app, has the ability to figure out what its citizens are searching online, what they’re saying online, what they’re buying online and where they’re traveling.
So there are a lot of these touch points in which the Chinese government could actually extract data from its citizens to just understand what’s going on on the ground and add to that mobile phone trackers. And right now, over the last two years of COVID, we’ve definitely seen the Chinese government improve in its ability to collect data on where your smartphone has been. So, for example, a lot of the COVID health codes that China has been using over the last 2 to 3 years, which have been tracking your position and where you’ve been going in the last two weeks just to ascertain if you’re a health risk. That’s an additional layer of data that the Chinese government now has the ability to collect on all its citizens in real time.
So I would say that China really stands out for the breadth of the data collection and its ambition. So it’s ambition really is the second reason why we see China as such a standout. Undoubtedly, China is probably the only country out there that hopes to use surveillance to create this techno-utopian state. And as we mentioned in our research and in the book, China has this ambition to use the data collected to analyze any future threats to its governance and to identify these threats quickly and do something about it, to create an alternative model to what democracy could offer. Because as a country itself, an authoritarian country does not have the same democratic institutions that developed Western countries do.
MICHAEL MORELL: Liza, that’s a tremendous amount of data you just talked about. And then earlier, you talked about the use of A.I. machine learning to do something with that data that would allow the Chinese government to take actions that they think are necessary to protect the state. Can you talk a little bit more, guys, about how they actually use AI? How they use machine learning? I think that would be insightful.
JOSH CHIN: I think probably the most vivid example in China and it’s one that the listeners of this podcast probably have heard of is Xinjiang, which is a remote region, far northwestern China on the doorstep of Central Asia. And Xinjiang stands out because it’s home to a large population of Turkic Muslims, including the Uyghurs, who have always chafed at Communist Party and at Han Chinese Rule. And so there’s always been a bit of conflict out there. And in recent years, the Communist Party, their solution to that has been to rule out what I think is probably the closest thing we’ve ever gotten to a sort of truly dystopian, sort of Orwellian surveillance state.And what they’ve done is they’ve gone out and, blanketed the entire region. It’s twice the size of Texas. It’s a massive place. But they blanketed the entire region in cameras and microphones, other sorts of sensors, security gates, to the point where if you’re a Uyghur moving around in Xinjiang outside or even inside your home, basically everything you’re doing is capable of being of being recorded and analyzed in real time.
And so all these sensors feed this data into a centralized data platform. It was actually built by a defense contractor. And it’s modeled on systems that the U.S. military uses to coordinate things, sort of complicated joint operations like counterinsurgency operations. And so all this data gets sucked into the centralized platform and then they use it there. So these cameras have, the sensors themselves have artificial intelligence built into them. So the cameras can recognize people’s faces. They can recognize people’s voices, that sort of thing. And then all the data gets analyzed on this platform in ways that it’s a bit of a black box. We don’t know exactly how they calculate it, but they use this data to categorize people according to the level of threat that they, the Communist Party perceives in that.
In one system that we looked at, there were three categories: unsafe, safe and average. And and so for people who were deemed unsafe and that means that they may have filled up their car with gasoline one too many times, or they may have a digital Koran on their phone, or they may be friends with the wrong people or whatever sort of biographical markers the Communist Party is using. The people who have those markers, who are seen as unsafe for a long time were being shipped off to a gulag of re-education centers.
MICHAEL MORELL: So it’s your discussion of Xinjiang in the book that really makes this whole thing chilling. But there’s a flip side of the coin, right? You tell another story about how these tools can actually be used for good. And then I’m wondering if you can talk about that a little bit.
LIZA LIN: Josh did a ton of the reporting in Xinjiang. On my part, I traveled to a lot of these wealthy Han Chinese cities that were located on the east coast of China and in those wealthy cities, what you saw were the same systems often built by the same manufacturers as well, that were used for repression in Xinjiang, were used to make city governance a lot more efficient. So, for example, the facial recognition that would possibly identify a security threat in Xinjiang would be used by the Chinese police to identify buckets of society that they deemed as people that could cause national instability or some harm. So people that the Chinese police would use facial recognition in the wealthier Chinese cities to identify would be ex-convicts, for example, or ex-convicts that were just released. And they wanted to monitor just to make sure that they were settling back into society much better. And then other people that they would use these facial recognition systems and CCTV cameras to identify would be fugitives on the run or people that they knew were notorious drug pushers.
So all these are the kind of people that you really, as a normal regular person, you wouldn’t be wanting to walk alongside on the street with. And it wasn’t just identifying people of interest as well. These same sensors that were sucking up video and car counts or a number of people counts were used to identify traffic accidents really quickly to get first responders to the scene. For example,we pull out a city in east China called Hangzhou and Hangzhou might not ring a bell to most of your listeners, but the companies that are headquartered in Hangzhou are probably would. So a company like Alibaba, for example, is headquartered in Hangzhou and a company like Hikvision, which is the world’s largest maker of CCTV cameras. They’re also headquartered in Hangzhou. So for these reasons, the Hangzhou government just has tended to really try and lean towards making Hangzhou a smart city in a digital sense. They’ve been very embracing of using technology to run the city, and using these companies as partners.
Hangzhou used to be the fifth most congested city in China, probably about two or three years ago. What the city government did was to put in place a system of CCTV cameras and sensors that were absorbing how many cars were on the street and using that to optimize traffic lights. So in periods of peak traffic, the traffic lights would stay green for cars a lot longer. And that kind of helped Hangzhou drop from fifth most congested to 57th most congested city in a matter of a couple of years. For a reporter that had been traveling in the region so frequently, I often went to Hangzhou for work. And if you were stuck in a Hangzhou jam, you would be stuck for 40 minutes and you’d be moving maybe a mile just because the Hangzhou Road Network was not established and the subway system was very immature and yet population and car population in particular probably doubled or tripled.
It was small things like these that the city government had been using these surveillance systems to help just make life more frictionless. I think the one example that really stood out to us in our research and reporting was when we found a guy whose mother had fallen into a river and they lived on the outskirts of Hangzhou City. But in that area, the Hangzhou government was still starting to put these systems in place. His mom had fallen into the river and he was really lucky. And they were very lucky because a neighbor was around the corner and fished her out. But she still was in a condition where she needed medical help. So the nearest hospital had sent an ambulance over to get her. And once she was in the ambulance, the driver had turned on this system that allowed traffic lights and sensors to recognize its license plate. So, you know, the same sort of computer vision that’s used in facial recognition in Xinjiang was used here to recognize the license plate of the ambulance and to turn all the traffic lights green from the point of picking her up to the hospital.
So that shaved off half the time that it needed for her to get to hospital and to get treatment and to drain out her lungs. So in such a life and death situation, it really makes you wonder if such surveillance systems actually might have a use. And in the case of Hangzhou, there were a lot of people that we spoke to which just felt like the positive externalities of having such a system outweighed any of the data collection, the more sinister data collection kind of outcomes.
MICHAEL MORELL: Given that, in Han China, is there a resistance in general to the surveillance state or is there an acceptance of it?
JOSH CHIN: This is actually one of the most fascinating aspects of doing all of this work. AndI think it’s our understanding that changed the most in looking at this. And I think we started off like a lot of people, with the assumption that in China privacy actually doesn’t matter that much. And in fact, if you look at the official Chinese government approved dictionary, the word for privacy doesn’t appear there until the late nineties. So it is a relatively new concept. And I should say, I think at least up until recently, a large number of people in China, the vast majority of them probably just didn’t really have the luxury to be thinking about things like privacy. They’re more concerned with putting food on the table. But in larger cities, as we did more and more work and talk to more people, we realized that in bigger cities with more educated populations you do have the beginnings of a privacy movement, an understanding of privacy.
And you had quite a few controversies actually involving big Chinese tech companies. One of them was Baidu, which is the Chinese equivalent of Google. In one instance, the CEO had been at a forum and said, we Chinese tech companies have a huge advantage because our customers don’t care about what we do with our data, so we can do anything we want with it. And that spurred a huge backlash. And what was interesting is the way that the Chinese government handled this was- it’s a country with this really extremely successful and powerful censorship apparatus. You would expect them to just use that to quash any talk of privacy. But actually what they what they did was a little bit of jujitsu where they actually threw their support behind a lot of these controversies and they had state media going after Chinese companies for mishandling customer data and essentially put themselves on the side of the people and kind of defined privacy concerns exclusively in terms of what private companies did.
So the government use of data has never really, at least again, not until recently, been scrutinized much in China. And now we’ve gotten to a really interesting period, which is, anyone who’s been paying attention to the headlines in China knows we recently had nationwide protests against the zero COVID regime that the Chinese leader, Xi Jinping, had been promoting. It’s hard to overstate how surprising these protests were. This was the biggest demonstration of public defiance against the Communist Party since 1989, since the pro-democracy protests back then. And at the center of it was this frustration with the controls that the Communist Party was imposing using its surveillance technologies. So I think we’re at a new moment. It’s be interesting to see how it plays out, because I think that this is one instance in which the Communist Party may have overplayed its hand.
MICHAEL MORELL: One of the things that I learned from the book was that China did not put this advanced surveillance system in place all by itself. It had some help. And can you talk about that?
LIZA LIN: Sure. So what we found in our research is that Silicon Valley has been there supporting the development of China’s surveillance state right from its incubation. China’s digital surveillance state, it’s in its current form, really only started to take root at the turn of the century. So around the year 2000, 2001, you saw the Chinese police a lot more willing to turn to digital technology to gather data. And when it did so, it ran these huge fairs, large expos in which it invited foreign companies to come down and to basically showcase their wares. Just to allow Chinese police to have a look at what’s out there and decide what they wanted to buy. And the companies that turned out at these exhibitions were all either pioneers of Silicon Valley or the who’s who of the digital world at that point. So you saw people like Sun Microsystems come down. You saw Cisco attend the exhibition and showcase what they had to offer. You saw the now defunct Canadian company Nortel Networks come down. And there was also Siemens from Germany.
And all these big companies were right there at the very first Chinese exhibitions trying to sell things such as fingerprint database collection systems or other censorship systems like filters for the great Internet Firewall. So we definitely saw Silicon Valley play a very important role in the incubation of the early Chinese digital surveillance state. Fast forward two decades to the present day. What you’re seeing is Silicon Valley companies, not not just providing systems, they’re actually providing components to the surveillance state.
So if you think about the hot dish drive industry globally, it’s really an oligopoly. The market share is mostly taken up by American companies such as Western Digital and Seagate. And these two companies have been known to be selling these hard disk drives to Chinese surveillance companies in large numbers and even optimizing some of the hard disk drives as surveillance hard drives for the Chinese. And it isn’t just like hard disk drives. You need the processing power as well. So chip companies like Nvidia and Intel, they’ve been known to sell their GPUs and CPUs these really high end chips to Chinese surveillance companies in order to power the data analysis that goes on on the back end.
If you think about Silicon Valley involvement, it’s huge. And not just in components. If you think about the financial aspect of things, a lot of U.S. private equity companies and venture capital companies were the first companies to give Chinese surveillance startups a leg up in the game by funding them. So there are a lot of ways that the West and in particular the U.S. has contributed to the development of China’s surveillance apparatus.
MICHAEL MORELL: I want to switch gears a little bit, guys, and talk about China’s export of its surveillance system to other countries. The conventional wisdom, as you know, says that China wants to shape the world in its image, in its authoritarian image as a surveillance state. And you push back on that notion. Can you talk about that?
JOSH CHIN: This view of China as wanting to push its system on the rest of the world, it makes sense. The people who have that view, it’s not coming out of nowhere. One of the pieces of evidence that people often cite is Xi Jinping himself, China’s leader. He’s talked about how China wants to make contributions to the human pursuit of better governance systems and that sort of thing. And it’s clearly diplomatically, it’s spending a lot of capital, a lot of time trying to cultivate relationships around the globe. In some cases trying to co-opt global institutions like the WTO and W.H.O. But what we saw when we started looking deep into this is just in terms through the lens of exports, of surveillance. And this model that they have in particular of this sort of digital authoritarianism is that it’s more complicated.
They are definitely selling these technologies and these systems and the ideas behind them around the globe. We don’t have really great data on this now. But the most recent number we have is from a scholar at the University of Texas named Sheena Greitens who dug around and found that there were at least 80 countries globally in the year 2020 who had imported state surveillance systems from China. And those were those range from authoritarian countries all the way to Western democracies, local police stations in France and Germany and that sort of thing.
One example that we saw in detail was in Uganda. Which is one of these countries that kind of hovers between the United States and China and maintains relationships with both. It’s nominally a democracy, but the leader there during the seventies won every election going back decades. And there are, of course, allegations of election fraud and whatnot. And China has been cultivating its relationship with Museveni for many years. And a few years ago, they sold him a sort of state surveillance starter kit. They flew Ugandan police out to China to go to Beijing, where they were given lessons in how these systems are used. And just like China, Museveni sort of initially said, oh, these systems are for fighting crime, but then quickly used them in political ways to track opposition political figures.
In some ways, what you saw in Uganda is China is exporting its system. But the way that it was used in Uganda is not the way it’s used in China. And that’s partly because it can’t be. China has all these advantages that a lot of other countries don’t. China has a lot of money. It has an immense, pretty well-trained, fairly tech savvy bureaucracy. A lot of countries really can’t replicate China’s model. And China, I think, realizes that. But what they’re interested in is promoting this idea of governments being able to use these technologies in whatever way they see fit. So the idea is not that they want to copy themselves around the world, but they just want to disrupt the existing Western dominated order which suggests that state power should be restrained and individual rights protected against invasions of this kind. They want to upset that conversation and make it okay for governments to use these technologies to control societies.
MICHAEL MORELL: I want to ask you guys about the West’s response to the rise of China’s surveillance state. What have we been doing? Have we been effective and are there better approaches to what we’re doing?
LIZA LIN: That’s a great question, because in the last couple of months, you’ve definitely seen governments such as the U.S. take action to stop the flow of the components that I talked about earlier and to the Chinese surveillance state. So on October 7th, you saw the White House put out new export sanctions. And with these sanctions, with these regulations, they barred U.S. companies from assisting in the development of China’s advanced chip and semiconductor industry. And that meant if you’re a U.S. person or a U.S. company, you couldn’t be selling components or you couldn’t even be working or helping in any form. U.S. persons were not allowed to help with the development.
One of the reasons cited was military. So they didn’t want to empower the Chinese military by helping with the development of high end chips. But they also wanted to stop the expansion of the Chinese surveillance state. And that was one of the reasons cited for these controls. So I definitely have seen some steps being taken in recent months. And on how these steps actually play out in China, it’s still a big question mark at this point. Because as with many export controls, you do see workarounds and in small quantities, Chinese companies still can get these high end chips in very small quantities.
So when you think about what governments are doing to stop the flow of either money or technology into the Chinese surveillance state, one company that often keeps coming up is a company called SenseTime. SenceTimes is China’s largest AI unicorn. And I would say it was probably one of the start ups that reading made this idea of facial recognition, using facial recognition in solving crimes really such a such a wide and global idea. So with Sensetime, what we saw was the U.S. government put one of Sensetime entities both on the entity list and on the investment blacklists. So if you were Sensetime’s Beijing entity, you couldn’t get chips from U.S. companies, you were on the entity list. And if you were Sensetime’s Hong Kong entity, then you couldn’t get any financing from U.S. investors.
And Sensetime is this company that had been really funded by people like Fidelity. Qualcomm had a stake in it. So what the U.S. government had essentially tried to do was to restrict the flow of money or chips to Sensetime. But in reality, because of all these loopholes that I mentioned earlier, it didn’t succeed. So with Sensetime, what we found out was when the company listed last year, 2020 in the Hong Kong Stock Exchange, it actually had a statement coming out saying that even though it was technically on the entity list, it was only one subsidiary of Sensetime that was on the entity list. It was its Beijing subsidiary. So its Shanghai subsidiary was free to buy whatever Intel and Nvidia chips that they needed to power the surveillance systems. So the loophole that I talked about earlier, one of these legal loopholes, just putting one subsidiary on the entity list wasn’t enough. And SenseTime was also put on a Treasury blacklist at the end of 2021, which really barred U.S. investors from investing in it. But what you saw when it eventually went to IPO was there was still U.S. money in it. And that was because, and this is according to a company, that was because its Hong Kong entity was on that Treasury blacklist. But it was a Cayman Islands entity that was listed in Hong Kong. So there are all these ways that the U.S. has been trying to stop the flow or stop its companies and investors from aiding the development of surveillance in China. But really, there are some limitations.
JOSH CHIN: The other plane on which this conversation exists. It’s kind of just on the level of policy and ideology actually. The real challenge for the U.S. completely apart from the technological question is the appeal of this model, of the Chinese model. Which I think is around the globe, it’s very simple. The Chinese government has this very simple message which is that these technologies can be used to increase security and improve certain aspects of lives in ways that we’ve never seen before. And they should just be, everyone should be able to use them however they want. And the U.S. doesn’t really have a good response to that.There is no coherent Western vision or democratic vision of how these technologies should be deployed.
And I think that’s, in part, in the U.S., it’s an issue. The U.S. has always been schizophrenic about the role of technology and data in society. Especially, since the rise of Silicon Valley, because Silicon Valley companies push this idea that it’s the free flow of information that gave rise to innovation in Silicon Valley. And so you can’t cut it off even though there are potentially bad side effects from having huge amounts of personal data floating around and being exploited by companies and governments. That’s just a price you have to pay for innovation. That’s the U.S. view of things. And the conversation hasn’t really moved significantly beyond that. It has started to move beyond that in other places. In Europe, for example, in the U.K., where you do have rules that are being implemented to really restrain some of the more invasive technologies like facial recognition. Right now, though, where you have regulators whose job it is to keep an eye on these things. But those are early days. And I think a lot of people feel that until the U.S. gets on board with it, you’re really not going to have a compelling democratic alternative to the Chinese vision.
MICHAEL MORELL: But we need one. We absolutely need one. The book is Surveillance State: Inside China’s Quest to Launch a New Era of Social Control. The authors are Josh Chin and Liza Lin. Josh, Liza thank you very much for joining us.
JOSH CHIN: It was a tremendous pleasure. Thank you, Michael.
