
Loading summary
Yachiu Wang
This is a given that whatever you say online, wherever you go, it's recorded. But the government is able to collect information. What are you doing online and then analyzing the information and make a prediction by analyzing the information to what extent you are being a threat to the government.
Michael Feinberg
It's the Lawfare podcast. I'm Lawfare senior editor Michael Feinberg, here today with Yachiu Wang of a human rights advocate who has studied the government of China for pretty much her entire career in the United States.
Yachiu Wang
I mean, Chinese AI is very powerful. It's collecting so much data and also processing so much data. It's going to make human rights activists resistant towards Chinese government much harder. But also, we need to keep in mind that the Chinese AI, Chinese surveillance, may not as successful as it appears to be precisely because China is not a free society.
Michael Feinberg
We are going to be talking about the development of artificial intelligence and scientific and technological exchange between the United States and China and what that says about both geopolitical concerns and human rights concerns. It would probably be helpful for many of our listeners who do not follow goings on within China at any real granular level to sort of understand what the landscape of the surveillance apparatus and efforts to control thought have been like before the advent of AI. Would you be willing to sort of walk us through what that general picture looks like?
Yachiu Wang
Well, I mean, I would divide it into two aspects. One is online. The other is offline online, as we all know. I mean, probably Americans know too, that a lot of websites were blocked in China. Facebook, X, Google, New York Times. I don't know whether Lawfare is being blocked or not, but it would be an honor if it's blocked. So there a lot of information you cannot access. But China has its own social media platforms. People can post things and discuss with each other, and then you cannot post certain words, certain events like the Tiananmen Massacre. You cannot talk about, you know, you can talk about cultural revolution, the Great Leap Forward, which are the terrible disasters done by the Communist Party. But there are, you know, things you can talk about. There are things that you cannot talk about. Then there's, you know, absolute forbidden topics like, you know, Xi Jinping's personal life, Xi Jinping's wealth. So there are the ways, you know, they censor what you can talk within the Chinese Internet. So that's basically how censorship works on the Chinese Internet before AI. Then there's the offline aspect. Of course, there were cameras right everywhere, even before the introduction of AI. But how the data can be processed and how the data can be used and synthesized and were much less effective and on a much less massive scale than before the arrival of AI.
Michael Feinberg
So let's talk about that. I mean, I think there was Washington Post story from a couple years ago where they walked through the video and physical surveillance apparatus in Beijing and the numbers were astounding. It was essentially if you're anywhere within the ring roads, you're going to be on camera non stop no matter what you're doing. Is that pretty much the case?
Yachiu Wang
Yes, I think, you know, there are different measurements and it's hard to say how many cameras are there in China in total, but I think there's one camera for every two or three people. That's the average. So you can just imagine the level of surveillance you're under.
Michael Feinberg
Okay, so that's talking about how the government collects information in terms of people's physical comings and goings. And the Internet surveillance and censorship that you mentioned, I'd imagine gives the government some insight into what people are doing online. But everything we've discussed so far is sort of means by which the government collects information once they have that information. How is it leveraged by the government and who is it leveraged against to the extent we would call it being weaponized or in a law enforcement manner?
Yachiu Wang
Well, let's, I mean, still let's talk about it in two aspects. When it's online, right, you say certain things online, you are able to post online and there are things you are not able to post online and the government is able to analyzing what you're saying online. Or you know, when you go to a website, it's real name registration. So obviously this is a given that whatever you say online, wherever you go, it's recorded. So but the government is able to collect information, what are you doing online and then analyzing the, the information and make a prediction by analyzing the information like to what extent you are being a threat to the government. So this is online, right? Then offline, the government is doing same that you know, have you, especially in the region of Xinjiang where the Uyghur population live, you know, where you go, who you visit, what are you doing in your daily lives, are all information that have been collected by the government. And you know, once it goes into the system, they are making a prediction in terms of to what extent your threat to, you know, to the society, to the government and the things they are collecting in Xinjiang has gone so ridiculous to the point I always use the example of, you know, they are collecting whether you are leaving your home from the front Door or the back door. So if you are living, for example, if you usually leave your home from the front door, but this time you're leaving home from the back backdoor, right. This information goes into the system and they are using it to make a determination as part of the data they are using to make a determination whether you're a threat to the government. Because the government is thinking, why are you suddenly leaving your home from the back door? There are other things like they're always tracking whether your cell phone has signals. If your cell phone somehow goes off track, right. Then the government starts to suspect why your cell phone doesn't have signal, why we can't. There are even things like if you buy weights from online, right. The government start to think why you're buying weights. Are you trying to exercise and make yourself stronger so you can participate in some kind of rebel groups? So all those data they are collecting get collected into a system and then some kind of algorithm that we have no idea privy to are making a determination whether you are threat to the government. And it's not like the government to feel your threat. And then the government go in and knock your door saying, you know, what are you doing? I want to know more. They, you know, they detain you, they sentence you based on that kind of data. So it's not just, you know, they're just checking on you and they're just watching on you. There are real consequences in places, especially in places like Xinjiang.
Michael Feinberg
Yeah. And one of the earlier seasons of Black Mirror, there was an episode where the protagonist was living in a world where everything she did online or socially created what I will refer to, and you will understand why I'm choosing these words, a social credit score that determined her tier of privilege and freedom within society. And it was taken as this sort of dystopian science fiction hour long entertainment by most people who saw the show. But I think to those of us who follow Chinese policy, we realized this is just the lived reality there.
Yachiu Wang
Yes, I mean, social credit score is a real thing, but it's in China right now. It's not as perfect as some people in the Western imagine. Like every movie is being tracked, every move is being factored into the government determination, the score you are and you know, the kind of category you're being put in. But there is this thing, it's less centralized, different municipal government have different systems. Even companies have that kind of systems. So this is for the mass population. It's real, but it's not working that perfectly as some people imagine. But on people like, you know, if you are dissident, if you're a human rights activist, if you are, you know, as I mentioned, a Uyghur in Xinjiang. So the kind of surveillance, a collection of data of your every move is very real.
Michael Feinberg
All right, so I want to sort of delve into something you've mentioned in your past. Two answers, and that's the Uyghur population largely living in Xinjiang. I think it's useful to talk about that first at a very general level, simply because I don't know that a lot of people outside the human rights and sinology community are really following what's going on there. There's a misconception, I think, among large parts of the west that China is in ethnically, racially and religiously, as well as linguistically homogeneous population, that everybody is Han ethnicity. Can you explain what exactly the Uyghur subpopulation is and why the Chinese government views them as a threat?
Yachiu Wang
The Uyghur population is a Turkic Muslim population in northwest China. There's about 13 millions of them. I mean, they used to have their own. They still have their own language, culture, religion practice. So you can get into the details, you know, to what extent the Chinese population control them in history. But it's undeniable. You know, they have their own religion, culture and language practices. So, you know, in the past, you know, the Chinese government has always have quite tight control over the population. But since 2016, the government really start to tighten the control over the population in response to some legitimate reality terrorist attacks carried out by the population. So in 2009, there was a massive violent attacks against the Han population in the area. Then in 2012, there were attacks in the train station in Kunming against the Han population. So there were sporadic terrorist attacks against the Han population. That was real. But then the Chinese governor came in response to that, had such incredible massive stringent crackdown on everybody in the region. So, you know, I think we needed to recognize both facts. There were some attacks, then the government had this completely disproportionate reaction to the attacks.
Michael Feinberg
Yeah. And now, I mean, we're looking at not just policing of the area. I mean, there have been documents smuggled out that talk about forced sterilization of the population, separation of children from parents, crackdowns on the use of the language. I mean, there are many scholars who have referred to it as an ethnic cleansing, essentially just not one that has progressed to the level of death camps yet. But it's interesting you mentioned the Uyghurs Also in the context of political dissonance. So I just. Before we move on to the next topic, is it fair to say that not all surveillance and oppression that might be happening in the mainland is geared solely towards ideology? That it's a little bit broader than that?
Yachiu Wang
I mean, in the Xinjiang region, it definitely targeting literally every person, every ethnic minority is not just, you know, the Uyghurs. There are other Turkish Muslim minorities too. So that is for sure. But on the Han population, so far, it mostly targets people the government considers, as you know, dissidents, human rights activists, troublemakers. But of course, the basic surveillance is still there against everyone. As I said, if you go online, anything you say, any activity you do, the government knows who is doing it and collecting the data. So I do think you're right. We need to differentiate the minorities with the Han population and the. Even among the Han population, there are the people the government considers as troublemakers, and then the regular population and the surveillance of them are different.
Michael Feinberg
And this was all in place well before the advent of what, in comparison to the past decade or two, we would call more advanced and sophisticated artificial intelligence models and procedures. Can you sort of talk about how the sort of explosion of knowledge about AI and its increased capabilities has changed this structure either for better or worse,
Yachiu Wang
let's say, you know, online. I mean, AI has been increasingly incorporated into the censorship, right? In the past, it would just, you know, there's this word you cannot post online. If you post your online, the government is able to identify, you know, you are the one who is doing this. Right? So now they are increasingly incorporating AI into the censorship itself. So it's more, you know, because people learned, you know, if I can't say July, June 4, because that's the day the Tiananmen massacre happened, I'm going to say, you know, May 34, which does not exist, but it's a reference to June 4. So people learned to circumvent the censorship in order to say what they want to say, you know, can reference Xi Jinping as emperor, right? But then because the introduction of AI really make that kind of circumvention hard because the AI is able to detecting, you know, your reference something else, right? This is one, the other, really. I think AI is being increasingly used in image detection because people know that if I post this word, it's can, you know, it won't past the censorship. So how about I just write on a paper and then post the picture on the paper? So the AI is being incorporated to detect such kind of Circumvention. So online, this is getting harder. I mean, offline, of course. I mean the camera, right? Facial recognition, they're able to identify your face through the camera. AI is in that. So there are also voice recognition, gate recognition. So there, there are ways that AI were incorporated in that kind of surveillance that the government can use to identify specific people even when your face is not showing. I mean, I don't know to what extent this gate recognition is working, is effective, but if you read the government materials, if you read tech companies materials, it's obviously there. They're saying we are able to invent this tool that able to recognize people's gate. So at least it's there.
Michael Feinberg
So there's a long history of people living under censorious regimes or authoritarian governments who, as you talked about it, find ways to communicate without being explicit. There's a professor from University of Chicago from many decades ago, Leo Strauss, who wrote about what he called the esoteric writing, wherein medieval philosophers would disguise their language to make points subtly that they couldn't overtly. And as we saw in Eastern Europe under the Soviets, almost certain cultural items take on a political cachet. Reading a certain book, patronizing a certain theater production become symbols or ways of expressing solidarity with people who are fighting against the regime in question. But it sounds like what you're saying is AI has now made the monitoring mechanism sophisticated enough that people living in China can't even do that sort of thing.
Yachiu Wang
I definitely think it's making it harder and harder. It's really getting hard. But I don't want to say that, you know, just make it impossible. Let me give you a very good example. During the COVID protests, I mean, China had this draconian control over the population. During COVID then, you know, people were protesting, then people went to the street. And as I said, the censorship is so severe. So, you know, what did they do? They held a white paper, right? A blank paper, as a way to protest. I mean, what the government can say, what are you protesting? I mean, I'm just holding a blank paper. So the censorship has gone so severe that under such circumstances, when you just hold a blank banner, a blank paper, people immediately know that you're protesting the government. The other good example is that the word Xi Jinping is so censored. Like, I mean, like as I said earlier said, you know, you can. People used to reference Xi Jinping as emperor, right? And then they're, you know, Winnie the Pooh, all that, that all got censored. So now it has gotten to the point that if you just say you him, people immediately start to think, you know, your reference Xi Jinping, him, the word him, the thing people, the first reaction people had towards this unknown him is Xi Jinping. So I mean, you cannot censor the word him because if you censor the word him, the Internet can't function. So in a way, you know, the government, such a draconian censorship is having an effect that every time when you say something that isn't referenced anybody, anything, people immediately think that it is a criticism of the government.
Michael Feinberg
Okay, so we've been talking largely about what happens within Chinese borders. Is the advent of this new technology and its availability to the Chinese government affecting populations outside the prc.
Yachiu Wang
I mean, I am part of the Chinese diaspora, I live in the United States. But of course I'm still subject to the censorship of the Chinese government first. You know, it's harder to post anything the government doesn't allow on the Chinese social media, just like the rest of the population inside China. And the other is that my communication with people in China are censored by the government. If I wanted to use social media platforms that are controlled by the Chinese government. I mean, because the Chinese government blocks Signal, WhatsApp and other platforms I used internationally, I have to use WeChat to chat with people in China. Right. Then that mechanism is censored. Of course, this is affecting the Chinese diaspora. Also the Chinese government's technology affecting people around the world. You know, there's this idea of smart city. Basically it's surveillance. Surveillance on the name of, you know, better governance. It's been exported around the world. Dozens of countries are importing Chinese surveillance equipment to, you know, so called, better govern, better manage their cities. So it's not just affecting the Chinese people inside, also affecting the Chinese diaspora, it's also affecting people around the world.
Michael Feinberg
Yeah. And if I remember correctly, the first major smart city project that was planned for North America was going to be in Canada and functionally built by Huawei.
Yachiu Wang
I don't exactly know whether it was in Canada, but it's true that, you know, they had the plan to, I mean, they have exported to dozens of, at least dozens of Belt Road Initiative countries, but they had a plan to, you know, exported to Western countries and cities too. But then there was a reaction to, which I think is a good thing, that reaction and awareness of Chinese government surveillance and its impact on human rights. So they were resistant to that, so that it was good.
Michael Feinberg
So it's interesting you mentioned the Belt and Road initiative though, because I Think most people don't realize there's also a cyberspace. I forget what the exact translation is, but there is a digital Silk Road initiative as well, where the Chinese government is essentially doing what it did with the larger Belt and Road Initiative, which is offering technology and services relating to cyberspace and relating to advanced technologies to third world countries at prices competitors cannot match.
Yachiu Wang
Yes, the digital circuit road is part of the BRI initiative, the Belt Road Initiative by President Xi Jinping and it's an important aspect of the bri. I think the government is increasingly emphasized that aspect of the bra, which is digital. And of course the government is setting all kinds of censorship and surveillance tools and know how to, to be right countries. And increasingly it's the, you know, the, the AI companies.
Michael Feinberg
Let's talk about that. Is most of the AI technology being used by the Chinese government being developed indigenously within China, or are they also making use of technology from other nations that in some circumstances might be more advanced than say, what DEEPSEA has come up with?
Yachiu Wang
Well, first of all, we should recognize a lot of the technology are indigenously invented by Chinese companies and the Chinese government has invested so much money to help those company to develop those tools. But of course, as we saw on the news that Anthropic recently has accused Deep Seq and other companies for the distillation or basically try to extract their development to benefit Deep Six development, I have no real understanding to what extent that's a legit or not legit stealing or just make use of what is available. I can't really say on that, but it's definitely part of it is indigenous, part of it's making use of what has been available internationally.
Michael Feinberg
It's interesting, and to be clear, I'm not an expert on AI technology, but I spent a long time studying how the Chinese government works with respect to its citizens and with respects to its expatriate or diaspora community. And there's a real problem that's not getting talked about a lot, which is to a certain extent authoritarian governments, if that's the proper word. I know there's debate over when that should be applied, but governments with less protections against state sponsored surveillance or state collection of data or what have you, they have a real advantage in the development of AI in as much as an AI system is going to be largely built on how much data it can ingest to learn from. And the fact is, in a country with less laws restricting the use of government surveillance, whether we're talking about the prc, whether we're talking about Russia, there's a dozen other nations we can name as well. They have a real advantage in terms of how much data and surveillance take and network analysis. They can feed into their AI engines to train them.
Yachiu Wang
I mean, that's absolutely true. I mean, just they can collect anything in China as long as the government wants. There's no way you can say, you know, this is my privacy, right? I don't want to give it to you. You can go to the court. The court is controlled by the government. So of course they collect anything that they want. And they are collecting anything they want by, you know, not just your face, your gait, your iris, your voice, anything. It's all biometric data. You know, people, some people try to sue the government. What can you do? You can, you know, suing government as an act of resistance, of course, and you get harassed or you can possibly be detained or imprisoned. So, yes, they have so much more capability in terms of collecting data, and there's just so much less resistance towards that kind of a government's collection of data in China, for sure. So the government, Chinese government definitely has an advantage.
Michael Feinberg
All right, so because China has so much more access to data, the tech companies in the United States very often argue that they need to be supported by the US Government specifically so that China does not get a technological edge. And it's usually argued about in the context of things like defense strategy, autonomous weapons, intelligence, analysis, things of that nature. But what we've been talking about is more of a human rights story so far. And to the extent that there are real human rights applications here, do you think the US Companies are engaged in those concerns, or should Americans fear as much from AI technology here as some Chinese citizens do overseas?
Yachiu Wang
Well, I think we should all fear AI companies. I mean, they're collecting this massive data. It's, you know, done in a way. There's no transparency and accountability so far. So in that regard, all AI companies have problems, and we need to be concerned about risk and then, you know, regulate those companies. But I think Chinese AI companies and American AI companies are different in a sense. First of all, the Chinese government controls Chinese AI companies and the US Government to much less standard. The US Government has control over American AI companies. Also, the Chinese government has authoritarian motives. Right? And America is a democracy. So in the way that, you know, there's debate, there's pushback, the American people can put on a check on the US Government that people in China cannot put a check on the Chinese government. So there are differences, there are similarities. We should all be concerned, but I still would differentiate Chinese AI companies and American AI companies also, especially concerning the large language models that Chinese companies are exporting around the world, the way that ChatGPT is everywhere around the world. But political censorship are embedded in Chinese AI companies that American AI companies don't because the government isn't telling ChatGPT, you cannot criticize the information that is critical of President Trump cannot be embedded in your AI system. It's not there yet. Maybe Trump wants that. But America is still a democracy. It's an open society. So in that way, Chinese AI companies, American AI companies are different.
Michael Feinberg
Yeah. And I firmly share that conviction and phrase the question provocatively to specifically to draw that out. But I think it does speak to a certain extent that, like, I don't know. And tell me how you feel about this. I don't know that AI is a malevolent technology in and of itself. And what we are seeing from China, though, is that you cannot evaluate that claim unless you look at the general legal regime outside the context of AI. So, in other words, if AI were to disappear tomorrow, it would not make China's government respectful of human rights, it would just make it less efficient at clamping down on them.
Yachiu Wang
Correct. I don't think AI is bad and good. It's how we design the product, how we use the product. What are the guardrails to protect people from harm by using the product. So, yeah, I don't think technology is good or bad. It hinges on how we use it.
Michael Feinberg
Okay, I want to sort of zoom out and talk about China tech and US Tech in a larger sense than just with respect to AI, there is a school of thought in the United States that for geopolitical reasons, and to a lesser extent for human rights reasons, which do not get nearly enough attention, there is a necessity for the United States to maintain technological superiority to the PRC government. Now, I'm going to put my cards on the table and confess like I was very much a part of that effort on the part of the United States, and I'm still 100% supportive of that effort. But you've lived in China. You've engaged with the government there in a way a lot of people have not, who are still free to speak. What's missing from that debate on the American side? As somebody who's lived in both countries and who has engaged with the repressive architecture in at least one of them, what do you wish more Americans were thinking about in this debate?
Yachiu Wang
Well, I think especially on the left in America, there's This less appreciation of just how brutal the Chinese government is, and also the lack of appreciation that the Chinese government is not just going to limit its brutality inside the country. When it can, it will export its authority, terrorism, human rights abuse. It is already doing it against the population outside of China. So in a way, I mean, at least prior to Trump administration, I so much more want America to maintain their superiority in technology and also other areas over China because I do see, fundamentally the two governments are different. America, with all its problems, is still broadly supportive of democracy and human rights and want that kind of governance model to be adopted around the world. So I do think America's hegemony played generally a positive role. But this has been challenged by the current Trump administration. But overall, I still believe and maintain that it's better that America is the hegemony than the Chinese government is the hegemony. But as we are seeing the evolution of the US Government's international actions, I don't know to what extent my belief will continue. But I still, as I see right now in the United States, there's such resistance to Trump's governance, whether it's within the country or its behavior outside of country. So I do hope that America can go back on track to be a country that is broadly supportive of human rights and democracy around the world.
Michael Feinberg
Yeah. So, okay, let's talk about that. Human rights and national security often get treated as very different things. I've never agreed with that particular dichotomy, and I know it's something you've spoken about before in other venues. Do you think there is a broad separation between national security and human rights advocacy, or do you see them more operating in parallel, or do you think they're the same thing?
Yachiu Wang
Well, I do think that, you know, America, for America to support human rights in China is good for Americans national security. You know, you want the Chinese, the country to be a stable country that is, you know, democratic, human rights respecting. So it facilitates, you know, a peaceful relationship with the United States. It facilitates international exchange, facilitates business. So it's good for China to be a human rights respecting country so you can have a normal relationship with China and you know, when for China to be more authoritarian, more repressive inside the country. The manifestation of that kind of repression inside the country is China oftentimes become more aggressive around the world, as we are seeing. Right. I mean, China has become more repressive in the past 10 years, also become more aggressive in the past 10 years. So I do think these two are very much related.
Michael Feinberg
So, all right. So let's put that in the context of everything we've been discussing earlier. There was a belief beginning largely with the first Bush administration and continuing very much into the Clinton administration, that if the United States was able to bring, able to sort of help shepherd an era of economic liberalism in the prc, that a political liberalism would inexorably follow. I think it's fair to say that it has not worked out that way so far.
Yachiu Wang
I agree. It's unfortunate. And I think people had that kind of expectation. Some of them, I think, were rationalization. Right? You wanted a lot of American businesses wanted to do business with China because it's good to make money, right? Then they start to rationalize by engaging with China, we're going to liberalize the country to make China more like America. So there were genuine belief that it was the right way to go. But there, I think were also rationalization for money making. But unfortunately it didn't happen.
Michael Feinberg
That idea that economic liberalization would lead to political liberalization has pretty much, at least over the past 35 years or so, proven not to be true. But now we get a similar argument that technological cooperation and scientific cooperation are the only way that humanity as a whole can progress. Because it's very common to hear that science doesn't operate within geographical boundaries. We build upon the steps that others have taken, we see where their experiments and hypotheses ended. We start from there, we incrementally move forward until eventually we get a great leap. Sorry, I should have used a different phrase talking about Chinese politics than great leap. But you understand what I'm saying and I'm going to phrase this in a very purposefully controversial way that I know a lot of people are going to disagree with. But does China's decision not to liberalize politically, but to actually become more reactionary as it became more of an economic powerhouse and more integrated into the world economy, indicate that we should have similar suspicions or not suspicions, but indicate that we should have a reluctance to integrate Chinese government sponsored research into the larger scientific community for the same fear?
Yachiu Wang
Well, I think, you know, I've been engaging this topic for a while. I feel the devil is in the details. I do think, you know, by integrating the Chinese scientific world into the international scientific world, you know, there's just so much collaboration going on. A lot of it was used by the Chinese government for malicious purposes. So I mean, you know, the governor will send Southern Talent Project, right Chair and Jihua, is that the right Thousand Talents program? Yes, yes. I mean, you know, people have a Collaboration with American universities on the name of some kind of a, for, let's say a public good. But the government does use the same, the results of the collaboration for military purposes, for example. So there's a large scale espionage. You probably know more than I do because you were in the U.S. government, so you had information. I don't. So that was a legitimate concern. But oftentimes collaboration are really just for public good, especially basic research. My concern is always that the US Government needs to know what kind of collaboration can be used by the Chinese government for malicious purposes and what kind of a collaboration is actually good and just for public, you know, for the public good, for that investment of science per se. So it really requires expertise within the United States government to be able to differentiate differences and, you know, make a good judgment.
Michael Feinberg
Yeah, there's an anecdote I tried to tell as many people as possible when I was supervising those types of espionage cases. And that is everybody who works in national security can easily think of a dozen famous cases where there was Chinese government sponsored theft of US technology. But there's a danger to a certain extent in focusing on that because the story they don't know is about the rocket scientist who was tenured at California Institute of Technology, who got so fed up with the suspicion and racism of the US Government in assuming he was going to do something wrong, which he never did and never had, that he was eventually driven from the United States, sort of softly persecuted, until he moved back to China and then developed their space program. So I just want to echo your statement that you do need smart people in this, the US Government who are able to differentiate between the sort of sharing we should be encouraging and the environment we should be fostering to welcome foreign scientists, but who also know that not everybody has the purest of intentions.
Yachiu Wang
Exactly. Yeah. I mean, that was already my concern even before the Trump administration. Right now my concern has become more grave because I known many talented people inside the government just can't stand their job anymore. So they left. So now I'm sure it's even, it's even a worse situation in their ability to differentiate what is legit and what is not legit. And also I have friends, Chinese friends, who are scientists who work in U.S. labs, U.S. universities. I mean, it's real, the pressure they are facing from the US Government sides, which the pressure they're feeling is from the administrations in their university, in their lab, but ultimately it's from the US Government. It makes them uncomfortable. They're automatically being suspected as a, you know, A spy. In fact, you know, they wanted to come to the US because it's a better country. It's a country that they can do real science in the environment that is more transparent, you know, that actually value their talent and they want to be in the U.S. they want to live in the U.S. they wanted their children to grow up in the U.S. because it's a better system, because it's a democracy, it's a human rights respecting country and they wanted to be part of it. But now they feel, you know, just by being virtually Chinese, I'm being a suspect. So you know, people do tell me that, you know, like I came here because it's a better country because I wanted to be an American. But now I just feel, feeling that, you know, America does want, does not want me. And then people feel hurt and how do they gonna react to that? Maybe some of them are going back to China, you know, not some, you know, a lot of them are going back to China because they don't feel welcome here. And then of course the Chinese government is very attuned to this and they are doing more to welcome people, go back to talents, to go back to China.
Michael Feinberg
So I want to tie this back to AI before we conclude and its effect on not just the US China rivalry, but also its effect on the Chinese populace both within the country and within the diaspora. If you were an American AI company and you're working on international ventures, or you're really an American tech company of any sort, you're Apple, you're Microsoft, you're OpenAI, you know, there's a million different companies we can think of and you have a lot of conflicting feelings about China as a market, as a partner, as a rival nation. What have you, knowing what you know about how the Chinese government leverages technology for, as you called it, malevolent ends, what do you wish American tech companies would think about before they approach China? Before as a market or a partner or a rival? That might not occur to them.
Yachiu Wang
You know, I feel America has been so successful as a country, American companies have been so successful, has so much to do with America as a democracy, as a human rights respecting country, as a society that people can speak freely without repercussion, without fear, right? So America has been able to thrive as a result of this system. So they need to keep that in mind. So when they interact with China, whether to see them as a partner, whether to see them as a rival, keep in mind that I'm good because I was born in a system that was open Transparent, human rights respecting and democratic. So you need to have confidence that promoting this is good, that it's in your advantage. Right. So coming from this place of that, this is good for me. This is why I'm successful and I wanted this to be brought around the world. It should be the fundamentals of how an American company should conduct itself, whether it's seeing China as a collaborator or China as a rivalry. Granted, Chinese AI is very powerful. It's collecting so much data and also processing so much data, it's going to make human rights activists resistant towards Chinese government much harder. But also we need to keep in mind that the Chinese AI, Chinese surveillance, may not as successful as it appears to be precisely because China is not a free society. You know, in the US as far as I know in New York City, there are a lot of surveillance, but there are, you know, civil society groups criticizing the surveillance of not being successful. I'm sure that in China it's happening, but there's no way that people can tell you the story. You know, you cannot read a story on the Chinese Internet from Chinese media saying, you know, the government's surveillance camera is not working for XYZ reasons. And the reason I can say that is because I have, you know, family members and friends in China, especially during the COVID and they were telling me, you know, this is the way I circumvented the health code. So the government didn't know that I actually went to that shop, actually I went to that city, which is prohibited at that time. So there are ways that citizens were doing all kinds of things to circumvent surveillance. And they are also finding out the government surveillance was not working as it claims, but they're not channeling it publicly because of censorship. So we just don't know to what extent it is working as the government claims it to work. So when we talk about surveillance and AI in China, we need to keep that in mind. That works not in favor of the Chinese government. I mean, the government is losing critical information in terms of how effective its control of the population is.
Michael Feinberg
And I think we will leave it there. Yachiu Wang, thank you very much for joining us today. It's been a pleasure from our end.
Yachiu Wang
Thank you so much.
Michael Feinberg
The Lawfare podcast is produced by the Lawfare Institute. If you want to support the show and listen ad free, you can become a Lawfare material supporter@lawfaremedia.org support supporters also get access to special events and other bonus content we don't share anywhere else. If you enjoy the podcast Please rate and review us wherever you listen. It really does help. And be sure to check out our other shows, including Rational Security, Allies, the Aftermath and Escalation. Our latest Lawfare Presents podcast series about the war in Ukraine. You can also find all of our written work@lawfairmedia.org the podcast is edited by Jen Pacha with audio engineering by Kara Schillen of Goat Rodeo. Our theme song is from Alibi Music, and as always, thanks for listening.
Date: April 8, 2026
Host: Michael Feinberg (Lawfare Senior Editor)
Guest: Yaqiu Wang (Human Rights Advocate, China Expert)
This episode features an in-depth conversation with Yaqiu Wang, a leading human rights advocate, on the evolution of surveillance and censorship within the People’s Republic of China (PRC), the expanding reach of artificial intelligence (AI) in state control, and the implications of these technologies both inside China and globally. The discussion also addresses the situation in Xinjiang, China's approach to the digital and global market, and the responsibilites and challenges for American tech companies interacting with the PRC.
Censorship Landscape Before AI
“You cannot talk about Tiananmen Massacre… So there are absolute forbidden topics like, you know, Xi Jinping’s personal life, Xi Jinping’s wealth.” (Yaqiu Wang, 01:44)
Explosion of Surveillance with AI
"Chinese AI is very powerful. It's collecting so much data... processing so much data. It's going to make human rights activists' resistance … much harder." (Yaqiu Wang, 00:36)
"They are collecting whether you are leaving your home from the front Door or the back door... why are you suddenly leaving your home from the back door?" (Yaqiu Wang, 05:24)
"It's not as perfect as some people imagine... but on people like dissidents or… Uyghurs in Xinjiang, the collection of your every move is very real." (Yaqiu Wang, 08:28)
“It’s undeniable. They have their own religion, culture and language practices. But since 2016, the government really start to tighten the control...” (Yaqiu Wang, 10:17)
“People learned to circumvent... but the introduction of AI really make that kind of circumvention hard because the AI is able to detect.” (Yaqiu Wang, 14:09)
“My communication with people in China are censored by the government… I have to use WeChat to chat with people in China… That mechanism is censored.” (Yaqiu Wang, 19:16)
“The government is increasingly [emphasizing] the digital aspect of the BRI… setting all kinds of censorship and surveillance tools and know-how.” (Yaqiu Wang, 22:01)
"A lot of the technology are indigenously invented by Chinese companies... but it's definitely part of it is indigenous, part of it's making use of what’s available internationally." (Yaqiu Wang, 22:54)
“Just they can collect anything in China as long as the government wants. There's no way you can say, you know, this is my privacy.” (Yaqiu Wang, 25:12)
"Chinese government controls Chinese AI companies and the US Government to much less [extent]... Also, the Chinese government has authoritarian motives. Right? And America is a democracy." (Yaqiu Wang, 27:08)
"Political censorship are embedded in Chinese AI companies that American AI companies don’t." (Yaqiu Wang, 27:08)
"I don't think AI is bad and good. It's how we design the product, how we use the product. What are the guardrails to protect people from harm." (Yaqiu Wang, 29:43)
"[T]he Chinese government is not just going to limit its brutality inside the country. When it can, it will export its authoritarianism, human rights abuse." (Yaqiu Wang, 31:25)
“For America to support human rights in China is good for American national security.” (Yaqiu Wang, 33:56)
"People had that kind of expectation. Some of them, I think, were rationalization... But unfortunately it didn't happen." (Yaqiu Wang, 35:35)
"My concern is always that the US Government needs to know what kind of collaboration can be used by the Chinese government for malicious purposes ... requires expertise ... to be able to differentiate differences and, you know, make a good judgment." (Yaqiu Wang, 37:55)
"Promoting this is good, it's in your advantage… This is why I’m successful and I wanted this to be brought around the world." (Yaqiu Wang, 44:10)
On the limits of censorship:
"If you just say 'him,' people immediately start to think, you know, your reference [is to] Xi Jinping… you cannot censor the word 'him' because if you censor the word 'him,' the internet can't function." (Yaqiu Wang, 18:39)
On the meaning of surveillance to China’s AI edge:
"They have a real advantage in terms of how much data and surveillance take and network analysis they can feed into their AI engines to train them." (Michael Feinberg, 24:30)
On the export of Chinese tech and AI:
"There’s this idea of smart city… surveillance in the name of, you know, better governance. It’s been exported around the world…" (Yaqiu Wang, 19:53)
On AI and resistance:
"People immediately know that you’re protesting the government… you cannot censor the word 'him'… so in a way, such draconian censorship is having the effect that anytime you say something ambiguous, it’s interpreted as protest." (Yaqiu Wang, 18:20)
| Timestamp | Segment Summary | |------------|-----------------------------------------------------------------------------------| | 00:36 | Power and limitations of Chinese AI and surveillance | | 01:44 | Censorship mechanics before AI in China (online and offline) | | 05:24 | Hyper-granular surveillance in Xinjiang (doors, cell signals, e-commerce) | | 08:28 | Reality of social credit scores and their decentralization | | 10:17 | Explanation of Uyghur identity and government crackdown | | 14:09 | How AI has begun to close loopholes for digital dissent and protest | | 17:15 | “White paper” protests and the political meaning of neutrality in language | | 19:16 | How Chinese surveillance affects diaspora and the global export of technology | | 22:01 | Digital Silk Road and surveillance tech as part of Belt & Road Initiative | | 25:12 | Data as a core advantage of authoritarian AI development | | 27:08 | Ethical differences between Chinese & American AI in censorship and oversight | | 33:56 | Linkage between national security and human rights advocacy | | 35:35 | Economic vs. political liberalization: lessons from China | | 37:55 | Nuanced perspective on scientific collaboration and the dangers of blanket suspicion | | 44:10 | Guidance for U.S. tech companies: hold fast to democratic values |
Yaqiu Wang emphasizes that while Chinese AI and surveillance are powerful and repressive—especially toward minorities and dissidents—the system is not infallible. Censorship is so intense that even the most innocuous forms of communication and protest can take on political resonance. Abroad, China’s tightening digital control affects both its diaspora and other nations importing its surveillance tools. Wang advocates for nuanced, informed American tech and policy responses, underscoring the continued value of American democratic principles as the foundation of both economic and technological strength.