---

For more than a decade, Jeremy Daum, a senior fellow at the Paul Tsai China Center at Yale Law School, has been a contributing editor at China Law Translate, a website he founded that hosts open-source English translations and analysis of Chinese laws. China Law Translate aims to expand accessibility and understanding of the Chinese legal system. Before the pandemic, Daum was based for many years in China, but has since returned to the U.S.. As part of his work, he has examined a wide variety of laws and regulations in China, including those pertaining to artificial intelligence.

Q: In recent years, there’s obviously been big developments in tech, particularly AI. How is China’s legal system adapting to those changes?

A: This is the issue of the day, and it’s something that I have followed closely. When I work in comparative law, I like to find issues that are affecting everyone and that every government has to grapple with, because it means we have something to talk about and can learn from each other.

China has been more proactive in many ways in addressing AI, but what they’ve done is taken their existing information controls and ported them. Their standard practice in tech is when there’s a new platform, or a new technology introduced, they say, “How does this impact our ability to keep imposing our previous stability controls?” And they find a way to do that.

They have focused primarily on generative AI. They do solicit feedback on regulations as they come out, and they modified the final published rules¹The Interim Measures for the Management of Generative Artificial Intelligence Services, which went into effect on August 15, 2023. because the original draft rules would have been restrictive to the point of hindering development of AI and its potential research uses. In the revision, they made it essentially so that the rules — the information controls — apply only to public-facing generative AI services, rather than also to internal uses of AI such as research and development.

The U.S. has gone in favor of allowing the new AI industry area to develop on its own, trying not to regulate it too much, though that is something China does as well — they try not to immediately stomp down on a new industry. They’ll stomp down on it once it’s a problem. In the U.S., it’s been largely, thus far, a profit-driven model, where we’ve had corporations, if they’re willing, sign on to voluntary agreements to accept certain controls, none of which have been put into law.

It’s not clear which system is more effective, because who knows if Chinese law is being actively implemented and enforced, versus these voluntary commitments which might end up being more effective, in some sense.

The Chinese regulations are still relatively new, but are you aware of how they have been working? Has it allowed for continued business development of AI, while achieving the protections that it aims to accomplish?

Seemingly. Like I said, the regulations — and there are various regulations out there — generally leave under-the-hood activity of AI development untouched.

In the U.S.-China forums that I’ve been in talking about AI, the two things I like to ask are, can we first define AI to make sure we’re talking about the same thing? Because I don’t think “AI” is that meaningful a concept technologically. At what point does something become AI? People don’t usually know. It’s more of a marketing label than anything else.

You see expressly written into laws [in China] that new areas will be less tightly regulated because we don’t want to crush them… That said, once something’s developed, we can see the ability of the government to completely crush an industry when it decides to. 

The second question I always like to put out there is, why the focus on generative AI? We see generative AI getting regulated a lot because we’re afraid of misinformation, and China’s worried about the ability for it to shift public opinion and for what they call “public mobilization”.

But the truth is, there are lots of cases where AI or high-level computing is useful that don’t really pertain to generative AI and misinformation — traffic pattern analysis, logistics, and things like this.

China has actually made some effort to register algorithms, and they do have a public-facing registry that tries to give a brief explanation of what each algorithm does. I think that’s ambitious, and it’s probably a good idea to let people understand what the functions are. It shows that you can get at what the public is concerned about without revealing source code or proprietary information. That sort of thing is proactive, and looking beyond generative AI is key.

You said that the Chinese government often allows room for innovation. Can you further describe this balance between allowing business innovation while still keeping some controls in place? And do you sense any changes to that balance, given the current economic slowdown?

You see expressly written into laws that new areas will be less tightly regulated because we don’t want to crush them. They make it clear that they are saying “Listen, we’re going to let it play out a little bit, and then we’re going to come in and regulate it. We don’t want to immediately regulate it.”

That said, once something’s developed, we can see the ability of the government to completely crush an industry when it decides to. An example is the after-school tutoring sector, where they just flat-out said this kind of business is not going to function anymore. That was a massive industry, and the businesses were wiped out without compensation.

So I don’t want to overstate the idea of them giving room for the development of AI. I’ve always marveled that businesses in China continue to function despite the uncertainty that, at any moment, they can just totally change their legal framework.

In the new economic situation, I have seen some loosening of restrictions. In particular there have been some recent changes to regulations limiting foreigners’ ability to operate certain kinds of businesses, like vocational training institutes. We’ve also seen relaxed visa restrictions for a lot of the world, including the U.S., to lure in more foreign tourism and investment dollars.

Beijing understands there is a need to get the domestic economy moving again, and that’s going to mean prioritizing that over some of the restrictions. But it probably won’t change things like restrictions on content or the ability for public mobilization, because those rules are solid, but it will allow a little more flexibility.

As you just said, content restrictions have been and will be a static feature of the Chinese system. Related to that, I’m interested in the end of the pandemic lockdown in China, including the White Paper Protests and that whole period, and how the Chinese legal system mobilized to suppress that dissent.

So much of it is a black box. That period is such an exceptional one. The Western reporting of it made it sound like it was 1989, like people were all over the streets, and it was a massive movement. It wasn’t actually that, but what was amazing about it was that, with so little protest tolerated, you had clearly coordinated, small-level incidents around the country, and that’s what generated a response. That’s what I was hearing from my Chinese colleagues, was that it was the coordination that was spooking the government badly at the end.

The impact of Covid everywhere, I don’t think it can be overstated. I think we really downplay it globally and in China. For years to come, we’re going to be understanding its impact.

My first trip back to China was right after reopening in 2023, and there was an attitude I saw of what I call learned hopelessness, where people were just defeated and they were angry, but almost too tired to be angry, like there was no reason to bother being angry because it wasn’t going to do anything. And that includes public sector employees, who had given up, just like saying, “I can’t even control my own government body, so what’s the point?”

When I went back a year later for a more extended visit, it was really different. A lot of the anger at the Party had dissipated. And while people certainly still had a resignation of being in a new world post-Covid, they were now saying, “Listen, as bad as things are, we in China are living a life better than some 80-plus percent of the world.” Of course, I have a subjective sample here, but it really was a noticeable shift.

What I saw wasn’t a nationalist resurgence, but it was pride in China, like they were saying, “You know, what? All in all, take the good, take the bad, we can do this, this is okay.” And I think that comes with growing awareness of what the rest of the world is like, including dissatisfaction with the so-called West. The West for years, since the financial crisis, has been losing some of its shine in China, because people are seeing the reality of it.

How do other countries, specifically the U.S., the West, Europe, compare on the AI regulation front?

It’s such a volatile area right now. Things are changing fast, but we’re going to see a lot of convergence, because the concerns are the same, and we’re going to have to try and get some regulatory norms.

China followed a lot of the principles of the European Union’s General Data Protection Regulation in creating their domestic rules for personal privacy, for apps, for collection of personal information. They require separate permissions for necessary information and anything beyond that, making it so that apps can’t deny you service if you refuse to provide non-necessary information. China’s always been very willing to look at what the rest of the world is doing.

Comparativists try to learn from looking at other models. That doesn’t mean you learn to see if someone’s doing it well. It means you learn from their experience, good and bad. I work in criminal justice, and while I don’t advocate the U.S. as a great model of criminal justice, we can certainly learn from what’s happening here.

When it comes to dealing with fake news and misinformation, we have a great thing to look at in China, in that they have censorship. They don’t have First Amendment concerns. But what we can do is we can look at China and say, “Okay, here’s a country that doesn’t worry about that. Let’s see if it worked.”

And what we always see is that China, with censorship, still struggles very much with these issues, with protection of minors, with having fake news. The term “fake news” was used in China before Trump ever said it, and it continues to be used for a reason.

Even in China, while censorship is very much a thing — I don’t want to downplay that — we’ve started to see government agencies realize that sometimes it is better to give a quick response to a breaking news story and try to satisfy the public with that, rather than use flat-out censorship.

You mentioned the EU’s data privacy law, and that China took inspiration from it. Can you talk about how China approaches privacy, specifically data privacy, versus how the West does?

So the U.S. — and here I’m just gonna say the U.S., because that’s what I’m familiar with, and Europe is quite different on privacy than the U.S. — we are very much founded on and committed to, culturally, the idea of limited government powers, and we talk about protection of our information from the government, such as with the Fourth Amendment’s limits on search and seizure.

China goes the other way. There aren’t so many restrictions on government collection of data, except those that the government imposes on itself. The government generally assumes that government employees collecting data will use it for the intended purpose only, and there is faith that the government will check itself. That’s a little bit of an overstatement, but that’s generally true, that there aren’t effective public controls.

…China, with censorship, still struggles very much with these issues, with protection of minors, with having fake news. The term “fake news” was used in China before Trump ever said it, and it continues to be used for a reason.

For corporations, China has come down much stronger than the U.S.. They are more able to directly regulate corporations and other citizens, and they have pretty firm rules about the collection of information and they listen to public concerns about that, because resolving disputes between civil entities bolsters the legitimacy of the party state.

If people can go to the party state, be it the courts or be it through policy making, and get a resolution to such concerns — for example, over the use of facial recognition to get into your neighborhood — if they can resolve that, it bolsters perceptions of their ability to solve problems, and keeps people from using non-legal mechanisms like protests to seek remedies. So they are more aggressive as a result.

It’s almost an inverse of the U.S. model. And this manifested itself long before big data was as prominent as it is now. I remember talking to a Chinese criminal justice scholar — my roots are in criminal procedure — who was saying that they really wanted private jails in China.

To me, as a liberal-minded American citizen, private jails are a horror. I said, “Why would you want that?” And she said, “Well, because when it’s the government, I don’t have any powers against it to find out about conditions and what’s going on.” But if it were a corporation under Chinese law, she would have more access to information and more realistic prospects for controlling and challenging their practice.

While we’re on the topic of data privacy, let’s talk about TikTok. Though the ban on the app is currently up in the air, does the law that led to the current situation mean other Chinese companies operating in the U.S. could be targeted next in some way?

The law does provide for banning other apps. While it does name TikTok, which is unusual in itself that it specifically names an app, it does also discuss other “foreign adversary controlled apps”. There’s a 1 million user threshold before any app can be forced to ban or divest.

I’m not a fan of banning apps. I lived many years in China dealing with bans on apps. I don’t like it there. I don’t like it here. I view the law and the discussion as, when can the U.S. government erect barriers to my accessing information?

There’s precedent in the U.S. that, rather than censorship, we counter harmful speech with more speech, and that even when something is unabashedly foreign propaganda, citizens have the right to access it. A labeling requirement was discussed as a less restrictive alternative, a way to alert Americans to the threat of misinformation and pro-China bias without having to actually ban or force a sale of the app. That, to me, makes sense.

In terms of data collection, I’m grateful that my government is finally looking at the protection of data. I just don’t know why you would look at it only in terms of foreign adversary-controlled apps when we have a global network of corporations buying and selling data. The data collection by companies versus data collection by authoritarian governments, to me, these are issues that shouldn’t be separated.

I was thinking, watching the whole fiasco, about Bill Clinton’s famous comment that China trying to crack down on the internet is like trying to nail Jell-O to the wall. These days that gets invoked a lot to show that we were naive, that China could nail Jell-O to a wall and control the internet.

My two thoughts on that are, one, China is constantly reeling from the new innovations and new manifestations of online life, putting out new regulations constantly because they can’t peg that Jell-O to the wall. And two, the naivete was in our thinking that we were somehow going to have an easier time because we were an open society.

Open societies, in many ways, are more susceptible to the changes in the information ecology. We’re a democracy in the U.S.. We depend, in theory, on the public being well informed, so misinformation has a much more direct impact on our political system. No comment on the results of that in our current situation, but I think it speaks for itself.

We’re seeing the whole world trying to make tracks to keep the Jell-O in at least, even if you can’t nail it into the wall. Can we make a funnel that at least keeps it as a contained substance?

Though it’s probably impossible to actually quantify, how much of the animosity between the U.S. and China — illustrated, perhaps, by the potential TikTok ban — can be attributed to misunderstandings specifically of each other’s legal systems?

That question speaks to the heart of what I view myself as doing. The U.S. spectacularly lacks understanding of China. 

Social credit is the best example of this — this belief that China has a citizen scoring system has made its way to the highest levels of the U.S. government. The EU and the U.N. have now both written rules to prevent the creation of a system that doesn’t exist. To keep thinking of it as something China is actively doing is just wrong.

There’s plenty of things wrong with China. Animosity, in many ways, is totally justified. There are freedoms that we hold dear and fundamental that simply are not protected in China. There is persecution of people, transnational repression, where people’s families that remain in China are sometimes approached and harassed in order to control the behavior of people abroad. These are all worthy of criticism, and I didn’t even mention Hong Kong and Xinjiang.

…there’s gross misunderstanding, or non-understanding of China. It is on the other side of the planet, and often gets used almost like a science fiction reality. We use it as if it weren’t a real place.

So how much of the animosity is legitimate versus due to misunderstanding? Yes, there’s gross misunderstanding, or non-understanding of China. It is on the other side of the planet, and often gets used almost like a science fiction reality. We use it as if it weren’t a real place.

We project anxieties and fears upon it, as with social credit — we have these growing concerns about data collection in the U.S., and we say, “What if the government was using data in an exploitative way?” And we say, “Well, that is happening in China,” even though that’s not what’s happening in China. China does have data concerns, but those aren’t them.

So, why is it a problem that people lack understanding of China? After all, let’s be honest, U.S. citizens lack an understanding of most parts of the world. So why is it a problem specifically with China?

It’s a problem because the two major powers right now are the U.S. and China. We have to deal with each other and address the problems. And if we’re addressing fake problems as much as we’re addressing real problems, it’s a waste of resources. It also grossly undermines our ability to address the real problems when we’re talking to China and we seem just as worried about something that is nonsensical as we are about something that’s real, like the treatment of Uyghurs in Xinjiang. It betrays our ignorance to the other side.

There’s also something that we used to call Marco Polo-itis, a phenomenon where a Westerner’s first visit to China doesn’t align with the horrors that they were expecting, and as a result, they over-correct, and they say, “Oh, China’s amazingly free. I was just the victim of a U.S. propaganda campaign.” And that’s not true either, right?

It almost feels naive in the current era of misinformation and tribalistic politics, but I think having an accurate understanding of the world is useful and helps you address things, and it helps you solve problems in a much more meaningful way.

---