Greetings from a world where sweatpants are forever…

…For the first time, we’ve had a net drop in subscribers in the past few weeks, so if you are willing and able to support keeping ChinAI content free for all, please please subscribe here to support ChinAI under a Guardian/Wikipedia-style tipping model (everyone gets the same content but those who can pay support access for all AND compensation for awesome ChinAI contributors). As always, the archive of all past issues is here.

Feature Translation: Douyu’s Jupiter, an open-source microservices framework

In Q2 of this year, people watched 5 billion hours of livestreams on Twitch, 56% more than Q1 (stay at home folks!). This is 2 billion hours more than Q2 from last year! Now here’s your weekly pop quiz: which Chinese company has more monthly active users (MAUs) than Twitch?

This week’s feature translation unpacks the tech stack of Douyu (斗鱼), the largest video live-streaming platform in China with 160 million monthly active users in 2019, about 20 million more than Twitch. *Not to be confused with the short video app Douyin (抖音), of which TikTok is the international version.

CONTEXT: Last month in June, Douyu open-sourced Jupiter, a microservices framework based on the Go language. OSChina, a very cool Wechat platform, talked with Lu Chao, an R&D engineer in Douyu’s Go team, to review Jupiter’s open source process as well as how Douyu’s technology stack has changed in recent years.

TECHNICAL TRANSLATION. This issue involves a lot of software development jargon, so let’s learn a little about: microservices:

• In contrast to a “monolithic architecture” in which a small update to one part of an app service requires all other parts of an app to adjust, a “microservices architecture” means that individual services have their own code and talk to each other through APIs (application programming interfaces). Douyu, for example, probably has one service that stores all the data on livestreams you’ve watched, one that deducts payments, one that manages Internet traffic and buffering, and another that uses algorithms to recommend livestreams.

  This approach was pioneered by Netflix, as Mayukhh Nair explains in a 2017 Medium post:

  Netflix literally ushered in a revolution around ten years ago by rewriting the applications that run the entire service to fit into a microservices architecture — which means that each application, or microservice’s code and resources are its very own. It will not share any of it with any other app by nature. And when two applications do need to talk to each other, they use an application programming interface (API) — a tightly-controlled set of rules that both programs can handle. Developers can now make many changes, small or huge, to each application as long as they ensure that it plays well with the API. And since the one program knows the other’s API properly, no change will break the exchange of information.

  The result: an ecosystem that allows a small team of operations engineers to quickly iterate changes in specific parts of the Netflix ecosystem without risking breakdowns to the entire service.

  

• Douyu initially adopted a Go-based microservices architecture at the end of 2016. Jupiter was built on top of this original framework. Go or Golang is an open-source language, developed by a team at Google, that helps programmers write flexible and modular applications. Per OSChina, Go is known as the “container language of the cloud native era.” (云原生时代的容器语言)

• The two mainstream microservices architectures on the market are Dubbo and Spring Cloud. Dubbo, a java-based RPC framework first open-sourced by Alibaba (interesting!) in 2011, was donated to Apache in 2017. Since then, the number of times it has been “Starred” on Github increased by 7429 in just 7 months (ranked 11th in Java category at Github). It’s also received recognition as a top open source software at Chinese conferences.

  • For more on what a RPC framework means, h/t to Girish Sastry, a researcher (policy) at OpenAI for sharing this explanation with me.

  • Also interesting: the list of companies that are using Dubbo are all heavy-hitter Chinese companies, including Alibaba, ICBC, and Didi Chuxing.

  • Jupiter is meant to be a complement to RPC frameworks like Dubbo. For instance, inside Douyu, the Java team uses the Dubbo framework and Jupiter provides a way for the Go-based applications to talk to the Dubbo-based services.

KEY TAKEAWAYS:

• Jupiter itself seems pretty impressive and has gotten a lot of feedback from contributors. The article notes that Jupiter’s Github has 1066 stars. Now, it has 2100 stars. When released, it was a trending Github project on June 3, 2020. One of the benefits of an OS project is a community to help maintain and improve code.

• Jupiter is also another indicator of the growing momentum for open source in China. We’ve previously covered a White Paper on China’s Development of AI Open Source Software in ChinAI #22.

• The diffusion of technology is not just technical, it’s organizational. Douyu started out using PHP and only switched to Go after rapid growth in the scale of the business, which required splitting up their engineering teams. Lu Chao cites Conway’s law to explain why Douyu switched: “organizational structure determines technical architecture.”

• Hey now, isn’t this a newsletter about AI? Well, microservices have been described as the operating system for AI, helping to solve problems with putting trained machine learning models into production: “When you have tens of thousands of model versions, each written in any mix of frameworks and exposed as REST API endpoints, and your users love to chain algorithms and run ensembles in parallel, how do you maintain a latency less than 20 ms on just a few servers?” A microservices architecture may be particularly suitable for increasing the scalability of AI applications, which often are composed of a multi-language stack in which data processing uses one language and modeling uses another.

FULL TRANSLATION: Douyu’s Jupiter, an open-source microservices framework

Reflection: US-China Relations and the “Security of Many Stories”

There is not a single English-language article about Douyu’s Jupiter. Yet I could convincingly make the case to you that stories like the one above about Douyu are just as important as the ones about Douyin’s forays with TikTok in the US. This gap in coverage relates to what I think is the most troubling trend in coverage of U.S.-China relations — what Chimamanda Adichie labeled years ago as “The Danger of a Single Story.”

Jeffrey Ding @jjding99

My patented 2-step guide to reading China-related news. 1) read. 2) ask yourself: how would I think about this development's effect on people if I didn't view everything through the lens of US-China great power competition. There it is, that's the secret ChinAI #93: Year 2 of ChinAIReflections on Technological Trajectories and the “China-watching” Communitychinai.substack.com

June 23rd 2020

3 Retweets35 Likes

Of course, aside from how the great power tech competition trajectory has sucked all the oxygen out of the room, there are a lot of reasons why stories about Douyu’s Jupiter don’t get told. One of these, to put it bluntly, is that US-China tech relations are being conducted and covered by lay people (like me) who don’t understand how technology actually works in key settings.

Consider this bizarre quote by American officials justifying Trump’s executive order on Wechat, taken at face value by two veteran NYT national security reporters:

WeChat is not widely used in the United States, except for by one key group — Chinese-born software engineers in Silicon Valley and other high-tech work forces, according to American officials. They use WeChat to collaborate on tough mathematical, software or engineering problems, trading solutions back and forth. Proprietary data can be scooped up by Chinese intelligence services, an American official said.

Let’s leave aside all the issues with the statement that “Chinese-born software engineers” are the key group that use Wechat in the United States. Actually, can I say one thing? For the love of God, can the U.S. newspaper of record please hire some people who have actually used WeChat or know people who use WeChat. And maybe also some people who drink bubble tea, if that’s not asking too much!

If these reporters and officials would have read this week’s feature translation or knew anything about how modern-day tech development worked, they would realize that you don’t trade solutions to tough engineering problems over Wechat. Much of it is done through open source platforms like Github, just like the Jupiter example covered above. It’s yet another example of the Trump administration’s 20th century approach to technology control. Again, as the Jupiter case shows, the tough part of adopting technological innovations is not so much in the blueprints or the IP; it’s the organizational and tacit knowledge.

More importantly, the danger of the single story of U.S.-China DECOUPLING! is that we get blinded to stories about how open source software could actually couple the U.S.-China tech ecosystems even more tightly together. The open source software movement, which is not slowing down anytime soon, is streamlining and standardizing software globalization. Open up a tutorial of the Go programming language, and you are greeted with:

“Hello 世界 (World)”

Each day hundreds of researchers and reporters think and write about AI solely in the context of two AI superpowers. On many days, I’d count myself among them. But how many stories get written about how neural machine translation could bring the U.S. and China even closer together? It’s not like we don’t have good evidence to ground these potential narratives. Econometric studies have proven that the introduction of machine translation systems has a significant effect on international trade.

Among “China-watchers” it feels like there are some resign themselves to the Trump admin’s “fuck China” strategy as a normal re-adjustment. “It is what it is”: This is a new Cold War, and we are destined to repeat history. But history is always being reshaped and changed. That’s why we can draw on a diversity of historical analogies for this current moment. And if history is what we make of it, why can’t the present and the future be just as malleable? Drawing inspiration from the TikTok teens, it is possible to create alternative memes. Why can’t the story of the U.S. and China as #coupledforlife be just as true as the one of #decoupling?

Jeffrey Ding @jjding99

China is the U.S.'s top collaborator on science & eng. publications; 26% of the U.S.'s int'l coauthored pubs w/ China; US is also China's top collaborator - 44% of China's int'l pubs w/ US coauthor. Per latest NSF report, 2018 figures ncses.nsf.gov/pubs/nsb20206 #coupledforlifePublications Output: U.S. Trends and International Comparisons | NSF - National Science FoundationThis report presents data on peer-reviewed S&E journals and conference proceedings reflecting the rapidly expanding volume of research activity, the involvement and scientific capabilities different countries, and the expanding research ecosystem demonstrated through international collaborations. Pu…ncses.nsf.gov

August 4th 2020

6 Retweets15 Likes

To draw on the feature translation about Jupiter one last time, we should be wary of monolithic architectures in U.S.-China relations. In ChinAI #90 I described one of my biggest worries as:

a neoconservative-stye co-option of U.S.-China policy. Basically, the Wolfowitz Doctrine but applied to U.S.-China relations instead of the Iraq War. A key component of that failure mode would be the rise of a small group of people who become the architects of U.S.-China policy and who all read the same narratives about US-China relations — they become insulated from the wider body of discourse and debate about U.S.-China relations.

This failure mode describes the Trump administration’s approach to TikTok, WeChat, and its general China strategy. What’s the solution? Neither Sway nor I have the answers, but one starting point would be to take a page from Jupiter’s notebook and strive for a distributed microservices approach to analyzing U.S.-China relations. In this ideal world, there would be a diversity of news organizations, blogs, think tanks, and individuals all telling different pieces of the U.S.-China story. It would be a polyglot (multiple programming languages) tech stack — some would focus on Chinese-language coverage, others on Japanese, etc. And we wouldn’t have arbitrary dichotomies separating China hawks and China doves, which entrench echo chambers. After all, a key part of microservices governance are the communication protocols that connect different code bases.

I’m truly hopeful we can avoid the danger of the single story and seek the security that comes from many stories. Call me naive. I can take it. But there’s a reason we will tell our kids fairy tales until the end times. Sometimes it’s okay to believe in a better 世界.

Thank you for reading and engaging.

*I’m behind on my reading so had to skip Four to Forward this week — will make up for it next week.

These are Jeff Ding's (sometimes) weekly translations of Chinese-language musings on AI and related topics. Jeff is a PhD candidate in International Relations at the University of Oxford and a researcher at the Center for the Governance of AI at Oxford’s Future of Humanity Institute.

Check out the archive of all past issues here & please subscribe here to support ChinAI under a Guardian/Wikipedia-style tipping model (everyone gets the same content but those who can pay for a subscription will support access for all).

Any suggestions or feedback? Let me know at chinainewsletter@gmail.com or on Twitter at @jjding99

Greetings from a land where Slow News exists…

…For the first time, we’ve had a net drop in subscribers in the past few weeks, so if you are willing and able to support keeping ChinAI content free for all, please please subscribe here to support ChinAI under a Guardian/Wikipedia-style tipping model (everyone gets the same content but those who can pay support access for all AND compensation for awesome ChinAI contributors). As always, the archive of all past issues is here.

Reflections: TikTok, Fast News, and the Technonationalism Dilemma

ChinAI is a newsletter in name only (NINO), and I intentionally try to stay away from the “fast news.” This is partly because of value-add/division-of-labor considerations — there’s no shortage of people who will go for the clout attached to the latest hot topic — but it’s mostly for personal wellbeing reasons. Chasing headlines adds so much stress and pressure to always be up-to-date and “first” on some issue. That’s why we’ve featured so much history on ChinAI lately. Slow news presents a nice check on the presentist bias we all suffer from.

This week I do want to tackle a fast news topic — Microsoft’s potential acquisition of TikTok — but in a slow news way. One good test to see if someone has taken the time to deliberatively think about the TikTok situation is if they can hold all five of these views at once:

via Alex Stamos, a researcher at the Stanford Internet Observatory.

It’s clear that the Trump administration is narrowly focused on one of these ovals — the risks of Chinese ownership. As with the Houston consulate decision and so many others, the Trumpian approach to China is to make fast news headlines rather than effective, sustainable policy. This critique is not new nor difficult to hammer home, and others have already laid out pathways for a more deliberative approach to TikTok.

Instead, in this issue I want to show how the Trumpian approach to China fuels a troubling trend I’ve pointed out in previous issues of ChinAI: the growing clout of techno-nationalist voices on the Chinese Internet. This week’s feature translation is of Luke Wen’s (卢克文) writing on the TikTok-Microsoft sale, which has racked up 100,000+ views in just a day. He’s one of the most prominent techno-nationalist voices, or what Ma Tianjie of Chublic Opinion has called “development bloggers” — an emerging, formidable force on the Chinese interwebs. Another rough heuristic of an article’s influence: As you can see from the screenshot below, 15 of my Wechat friends read his blog post on this issue on the TikTok-Microsoft sale. ***Very rarely do I see more than 10 of my friends having read a particular article when I scroll through my subscriptions):

Feature Translation: Why is TikTok being forced to sell to Microsoft?

After working in e-commerce operations for a decade, Luke Wen set up a WeChat public account in March 2019 that has gained millions of readers. Aiming to provide “the most in-depth breakdown of international politics,” the account frequently features writings on technological competition and great power relations. Often writing in a genre I label “techlore,” his style involves generalizations about the fates of nations from personal anecdotes and dramatic stories.

• On why the U.S. is so opposed to TikTok: “Since the United States is powerful, it has always been the master of martial arts, with many younger disciples. Therefore, if there is a public opinion showdown in the world, we have no chance of winning at all, just like the Chinese Navy if it encounters the US Navy in the depths of the Pacific Ocean. But around China’s territory, we still have a certain chance of winning a fight with the United States, and this is also true for the war over the mobile Internet.

  The international public opinion platform is the Pacific Ocean where the United States has the discourse power, revealing the soft power of the United States. It has always been a place where the United States’ mouthpiece is invincible. However, a Chinese nuclear submarine has suddenly charged in, destroying everything it sees. 

  Tik-Tok is this nuclear submarine.”

Playful slang is peppered throughout the pieces, attempting to appeal to netizens. In this piece, for instance, he uses the disparaging nickname 蓬胖 (Fat Peng) to refer to Sec. of State Pompeo (who’s name is usually translated as 蓬佩奥 (Peng Pei Ao ).

• He writes, “I beg you Pompeo, can you show me some evidence? If there is no evidence, I have to sue you for libel. (But this guy has a thick-skinned face, so it's probably useless to sue.)”

My aim here is not to give Luke Wen more of a platform. This would be akin to what people do when they “debate” the Global Times editor to get more Twitter points. Many Chinese commentators have denounced Luke Wen’s posts for lack of systematic research, sensationalism, and even making up stories. Rather, my aim here is to present two key takeaways:

1. Reading the full translation gives you a sense of the one-dimensional-thinking of these techno-nationalist “development-bloggers.” I’ve commented in some fact-checks and flaws in the thinking. And sadly, it also shows how many of the same tropes are mirrored in the logic of the Trump administration and its enablers.

2. It highlights the dangers of a technonationalism spiral. Is there anything that empowers the impoverished thinking of hyper-nationalists more than the ability to point to the writings and actions of hyper-nationalists in another country? Let me be clear: there are still a lot of diverse voices on the Chinese Internet that hold five-oval, multi-dimensional views about issues like TikTok. In this post, Luke Wen even positions himself against all of China’s public intellectuals (公知): China’s public intellectuals serve as fanatical missionaries of universal values ​​in the West, taking on the "heavy responsibility of awakening the foolish people.”

   But there’s a risk that Trump’s “fuck China” strategy lends further legitimacy to voices like Luke Wen’s. Back in May of 2018, Ma Tianjie wrote in a Chublic opinion blog post, “If Donald Trump’s trade war has any effects, one of them would be uniting the Chinese internet under the flag of industrial self-armament.” To borrow the words of Cardi B, “it’s not a threat, it’s a warning.” Be careful — I plea.

FULL TRANSLATION: Why is Tik-Tok Being Forced to Sell to Microsoft

ChinAI (Four to Forward)

Must-follow: all these people

ChinAI provides a biased, limited cross-section of the ChinAI space. To develop a multidimensional view, get a balanced diet of all these journalists and researchers who've spent years undertaking sharp & difficult work on China plus AI & tech — including Christina Larson herself.

Christina Larson 可心 @larsonchristina

Here are some journalists and researchers who've spent years undertaking sharp & difficult work on China plus AI & tech: @YuanfenYang @LiYuan6 @gwbstr @joshchin @lizalinwsj @dakekang @EBKania @MaraHvistendahl @mattsheehan88 @GaoYuan86 @_KarenHao @Kate_OKeeffe @suilee @suelinwong

July 29th 2020

32 Retweets100 Likes

Should-read: Sheena Greitens’ Critique of Ross Anderson’s Atlantic Piece

Sheena Greitens expertly takes apart this piece. I won’t say more other than I remember leaving a get-together with friends to talk to Ross about this piece. Let’s just say there’s a reason none of the nuance I tried to add made it into the story. He clearly had a frame he was already committed to writing within the bounds of.

Sheena Greitens @SheenaGreitens

Was initially excited to see this piece on China's AI & tech surveillance from @Anderson @TheAtlantic. But it too glibly repeats a lot of China's own hype about surveillance systems w/o enough tough scrutiny. And it needed better fact-checking. Thread: theatlantic.com/magazine/archi… 1/nThe Panopticon Is Already HereXi Jinping is using artificial intelligence to enhance his government’s totalitarian control—and he’s exporting this technology to regimes around the globe.theatlantic.com

July 29th 2020

122 Retweets281 Likes

Should-read: A New Wave of Podcasts Shaking up Chinese-Language Media

Shen Lu in Politico notes a new wave of Chinese language podcasts and newsletters — positioned in-between Chinese state media control and American-centric Western media. These include: In-Betweenness, Loud Murmurs, News Lab (a newsletter run by Kecheng Fang, a journalism professor based in Hong Kong.

Should-read: National Security, Antitrust, and AI

How are national security considerations likely to affect US antitrust decisions on AI? Cullen O’Keefe investigates the question by looking at historical cases in a new FHI technical report.

Thank you for reading and engaging.

These are Jeff Ding's (sometimes) weekly translations of Chinese-language musings on AI and related topics. Jeff is a PhD candidate in International Relations at the University of Oxford and a researcher at the Center for the Governance of AI at Oxford’s Future of Humanity Institute.

Check out the archive of all past issues here & please subscribe here to support ChinAI under a Guardian/Wikipedia-style tipping model (everyone gets the same content but those who can pay for a subscription will support access for all).

Any suggestions or feedback? Let me know at chinainewsletter@gmail.com or on Twitter at @jjding99

Greetings from a land where 2.5 generations ago there was astonishingly widespread belief in world government…

…as always, the archive of all past issues is here and please please subscribe here to support ChinAI under a Guardian/Wikipedia-style tipping model (everyone gets the same content but those who can pay support access for all AND compensation for awesome ChinAI contributors).

Feature Translation: Tencent’s 2020 White Paper on AI

This week’s joint translation is with Caroline Meinhardt, a GovAI summer fellow, who initially discovered this white paper. In last week’s issue, she offered four takeaways from the World AI Conference in Shanghai. Make sure to read her other work for MERICs on China’s “new infrastructure” concept and how China is turning to European suppliers to cut its reliance on US technology as well as the hidden challenges of China’s booming medical AI market for China Business Review.

CONTEXT: Introduced on July 10 at the WAIC, the White Paper provides a glimpse at how Tencent is thinking about AI. Five chapters on the macro-environment, technical research, AI applications in combatting Covid, new innovations for the economy, and institutional safeguards. We’ve translated the excerpts we found the most interesting, but the full translation gives the entire table of contents.

Our KEY TAKEAWAYS:

• Preface and Chapter 1, The Macro Picture: “ubiquitous AI.” Si Xiao, head of Tencent’s Research Institute, describes"ubiquitous" in the first sense in infrastructure construction. “Guided by the favoring winds of ‘new infrastructure,’ artificial intelligence technology will gradually transform into fundamental service installations like the Internet and electricity.” “Ubiquitous" in the second sense in more diverse application scenarios and larger audiences (adoption sectors).

• Chapter 2, directions for technical research: A) in general ML — increase the efficiency of small-sample learning && develop “offline reinforcement learning” (RL algorithms that learn from a fixed batch of data without exploration); B) in computer vision — enhance defense against adversarial attacks && deepfake recognition and counterfeit-detection technology to curb AI abuse

• Deep synthesis chapter (4.3): chapter overwhelmingly focuses on the positive applications of what it calls deepfake synthesis. Tells a kind of evolution story from the original crude applications (e.g. deepnudes) to more and more creative, innovative, and socially impactful applications across entertainment, e-commerce, content creation etc --> the phrase “emerging from the shadow of pornographic face swap videos to usher in an era of commercialization” is used in both the first and last paragraph. Examples of deepfake synthesis include an ALS patient speech synthesis program to an interactive Martin Luther King exhibition.

• AI for FEW (Food, Energy, and the Environment) chapter (4.5): Interesting details on Tencent’s agriculture AI projects and how they actually improve efficiency — includes an example of collaboration with a European university (in the Netherlands). This seems to be somewhat of a priority for Tencent.

• Chapter 5, Institutional safeguards: Argues that establishing a multi-level governance system is key but cautions against hasty regulation that’s too strict/inflexible. Seems to us that Tencent is being unusually blunt about discouraging strict regulation — they’re not even pretending. Also emphasis on China’s strong system of pilot zones that provide the space for cutting-edge development/experimentation. Calls for int’l cooperation and for the tech industry to shift from its current “technocentric” model to a technohumanitarian collaboration model (技术人文协作模式).”

FULL TRANSLATION: Tencent 2020 White Paper on AI

ChinAI (Four to Forward)

Must-read: Intermingled (State and Private Companies) Censorship of the 19th National Communist Party Congress on WeChat

For The China Quarterly, Lotus Ruan and a team at Citizen Lab shine a light on how censorship actually gets implemented on company platforms:

lotus @lotus_ruan

Just out in @chinaquarterly: my @citizenlab colleagues and I examine the relationship between political events and info control on WeChat thru a longitudinal analysis of keyword censorship related to China's 19th National Communist Party Congress (NCPC19) The Intermingling of State and Private Companies: Analysing Censorship of the 19th National Communist Party Congress on WeChat | The China …The Intermingling of State and Private Companies: Analysing Censorship of the 19th National Communist Party Congress on WeChat - Lotus Ruan, Masashi Crete-Nishihata, Jeffrey Knockel, Ruohan Xiong, Jakub Dalekcambridge.org

July 2nd 2020

56 Retweets93 Likes

Should-read: Image “Cloaking” for Personal Privacy

The SAND Lab at UChicago released a new algorithm and software tool called the "Fawkes" project. It provides an “image cloaking” service to protect against unregulated 3rd party facial recognition services (e.g. clearview.ai, “has now downloaded over 3 billion photos of people from the Internet and social media, using them to build facial recognition models for millions of citizens without their knowledge or permission.” Link contains a very informative FAQ page for those who want to quickly digest key points.

Should-read: Banning TikTok is a Terrible Idea

For SupChina, Samm Sacks lays out a more sensible approach for balancing data security/censorship issues with the U.S.’s broader goal to “offer an alternative to Chinese cyber sovereignty with a vision for internet governance to help better secure data online and prevent the spread of disinformation. These are problems that are bigger than TikTok and must be dealt with in a separate lane from the U.S.-China conflict. Let’s leave the creation of national walls in cyberspace to Beijing.”

Should-read: Messier than Oil:
Measuring Data Advantage in Military AI

CSET issue brief abstract: “Both China and the United States seek to develop military applications enabled by artificial intelligence. Will China be able to accelerate this effort thanks to its access to vast quantities of private-sector data and robust surveillance? …in assessing whether the United States or China has a “data advantage” in the military AI realm. Authors Husanjot Chahal, Ryan Fedasiuk and Carrick Flynn find that commercial data, while useful, will prove less relevant for military operational AI than generally thought, and that emerging technical approaches might reduce the role of big data in AI competitiveness.”

Thank you for reading and engaging.

These are Jeff Ding's (sometimes) weekly translations of Chinese-language musings on AI and related topics. Jeff is a PhD candidate in International Relations at the University of Oxford and a researcher at the Center for the Governance of AI at Oxford’s Future of Humanity Institute.

Check out the archive of all past issues here & please subscribe here to support ChinAI under a Guardian/Wikipedia-style tipping model (everyone gets the same content but those who can pay for a subscription will support access for all).

Any suggestions or feedback? Let me know at chinainewsletter@gmail.com or on Twitter at @jjding99

Greetings from a land where we are here to keep watch, not to keep…

A full plate this week.

• Appetizer: completed tech neutrality translation.

• Main course: takeaways from the recent World AI Conference (WAIC) in Shanghai, by Caroline Meinhardt.

• Dessert: Some ramblings on analogies and race in U.S.-China relations. I say some things that people aren’t willing to say out loud at the end, so if you stick with me until the end of the meal, I can promise it’s worth it, but I can’t promise it won’t unsettle your stomach.

…if you like the meal please leave a tip and subscribe here to support ChinAI under a Guardian/Wikipedia-style model (everyone gets the same content but those who can pay support access for all AND compensation for awesome ChinAI contributors). As always, the archive of all past issues is here.

Feature Translation: Demise of Tech Neutrality (Completed)

Finished up the last parts of this lengthy essay. A few quick-hitters in the new stuff:

• Part 5: Covers three cases: 1) Qvod and 2) Bytedance’s Toutiao were both punished for spreading pornographic and obscene content. In 2016, Wang Xin, head of Qvod (which had 80 percent of China’s online video-streaming market at the time), was sentenced to 3.5 years in prison. This “announced the low tide of technology neutrality in the Internet age.” 3) He Jiankui’s bioethics violations.

• Part 6: Author’s explains why technology neutrality is complicated by the view of technology as a process vs. technology as a thing: Modern people increasingly feel that each technology must be viewed as a "process" rather than a "thing.” We care about the application scenarios of AI, and we care even more about how it is implemented in algorithms; we care about the benefits brought by gene editing, and we care even more about whether it has potential risks in biochemical mechanisms. We have conjured up the courage to open the "black box" and exposed the dark side that had long been concealed by the myth of "technology neutrality.”

• Part 7. It concludes: Where does humanity’s view of technology go? If pure "technology neutrality" is merely a fantasy of technological optimism, will its demise trap us in extremely pessimistic relativism or even nihilism? Since all technological processes carry human likes and dislikes and even various structural inequalities, is it possible to declare that any unbiased universal technology is a bust? Are we destined to only choose between "white AI" or "black AI", “male technology" or "female technology"? Regardless of the way out, at least one thing can be determined: technology neutralists…must make efforts to adapt to the times. After all, we cannot go back to the era of "technology neutrality" as a first principle.

4 Takeaways from WAIC 2020 in Shanghai

*Thanks to Caroline Meinhardt, an analyst at MERICs, who we’re fortunate to have as a GovAI fellow for the summer, for taking notes on the World AI Conference held in Shanghai earlier this month. Her takeaways follow:

Context: From July 9-11, Shanghai held its third annual World Artificial Intelligence Conference (WAIC) with the theme ‘Intelligent Connectivity, Indivisible Community.’ For the first time, the entire agenda of keynote speeches and panel discussions was streamed online due to Covid-19. Organized by the Shanghai municipal government, the WAIC brings together the who’s who of Chinese (and some international) AI leaders from government, academia and industry every year to discuss the trajectory of AI development, explore the latest AI application trends and present new AI solutions (and gimmicks). 

The ‘responsible development’ of AI featured prominently during this year’s conference, seemingly confirming speculation that 2020 will become the year of AI governance in China. In a 3-hour AI Governance Forum, more than 30 Chinese and international speakers discussed the ethical and safety challenges of AI development and opportunities for global cooperation on these issues. As such, the Forum provides an interesting window into the key trends that are shaping China’s approach to AI governance:

1. Turning principles into action: A common thread throughout most speeches was the need for China to move away from high-level AI governance rhetoric and towards more nuanced discussions of specific scenarios and concrete solutions. To that end, speakers provided detailed insights into their ongoing research on AI security issues and ideas for mechanisms to assess the trustworthiness of AI systems. Particularly noteworthy were the formal announcements of two new Shanghai initiatives to strengthen AI governance: 

• The Expert Advisory Committee of the Shanghai AI Pilot Zone has created a new Governance Working Group that will oversee the implementation of AI governance principles on the ground. Made up of scientists, legal experts, ethics experts and company representatives, the Working Group will lead AI governance research, make recommendations on AI legislation and principles, and promote international collaboration.

• The Committee also published a new document that aims to further advance AI governance efforts in Shanghai, entitled ‘Action Proposals on Collaborative Implementation of AI Governance Principles.’ The document is the first to outline concrete steps for implementing the national AI governance principles issued by the Ministry of Science and Technology’s AI governance advisory committee in 2019.

2. Chinese companies play an active role in AI governance: Xue Lan (dean of Schwarzman College) and other speakers emphasized that Chinese companies should, and indeed already do, take concrete action to address AI ethics issues. CEO and co-founder of Megvii, Yin Qi, made an effort to demonstrate exactly that during his speech: He presented several concrete measures his company is taking that have been developed and executed by internal working groups across R&D, product development, customer relations and operations. Such measures include the application of federated learning techniques to ensure data privacy protection and the creation of instruction pamphlets for customers and partners that outline Megvii’s standards for the ethical usage of its AI products. Yin Qi’s speech and its prominence at the Forum highlighted that the company has emerged as an industry leader for corporate action on AI governance. Megvii claims to have been the first Chinese company to establish an AI ethics committee and most recently set up a dedicated AI governance research institute.

3. Global cooperation is key: Another common theme was the emphasis on international exchange and cooperation which is essential to create a global AI governance system. While the calls for cooperation were vague, it is worth noting that several speakers emphasized that Chinese AI ethics principles are aligned with the many international principles that have emerged in recent years. The AI Governance Forum itself also says something about China’s vision: It sees its companies and research institutes as playing an important role in shaping international AI governance efforts, but also sees itself as an influential convener of the international AI governance community.

4. What wasn’t said: Unsurprisingly, the human rights concerns surrounding the use of facial recognition technologies for state surveillance went unmentioned. Discussions of ethical issues such as algorithmic discrimination exclusively focused on the ability of companies to limit consumers’ options and cited examples from abroad, such as instances of racial discrimination in the US. 

On analogies, U.S.-China Relations, race, and falling in love

Part 1: A New Analogy Awakens

Instead of a ChinAI Links (Four to Forward) this week, I want to riff a little bit about Julian Gewirtz’s recent must-read Twitter thread on analogies for U.S.-China relations that are more creative alternatives to a new “Cold War.” He points out that the Sino-Soviet split, which crystallized in the early 1960s, offers many interesting parallels. Like the U.S.-Sino relationship today, China also benefited from its relationship with the Soviet Union in terms of technology flows. Shifts in China’s domestic politics back in the 1950s, like Xi’s consolidation of power in recent years, presaged the split.

I’ve been quite outspoken about the flaws and dangers of the “Cold War” analogy, but Julian’s thread was a wake-up call for me. It’s easy to tear stuff down and endlessly critique something. It’s much harder to construct a better alternative. It’s especially hard because no analogy is perfect. As Julian notes, the Sino-Soviet relationship of the 1950s, in contrast to the U.S.-China relationship today, was based on shared communist ideals. I think the more fundamental difference is that the PRC wasn’t just splitting from the Soviet Union but that it was also, in some respects, turning toward the U.S. — part of a “strategic triangle.” The Soviet Union and China also shared a territorial border. This list of differences can go on.

Indeed, on a deeper level, language of all forms — including analogical comparisons — can never fully capture the essence of what we want to describe. We can never fully describe what it feels like to “fall in love.” To even get close might require a book-length treatment, or scores of tumblr posts. When faced with limits on words and time, we reach for analogies. Falling in love is like giving oneself over to the force of gravity. To fall in love.

Part II: Analogies Strike Back

This isn’t just fanciful wordplay. Imperfect analogies come with real costs. For instance, the historical analogy to Munich — a callback to the failures of appeasement before World War II — played a meaningful role in policymakers’ decisions to intervene in Vietnam.

So, what gives? Should we just avoid historical analogies like the Sino-Soviet split or the new Cold War altogether? Abstinence is not the solution. In their book, Thinking in Time: The Uses of History for Decision-Makers, Ernest May and Richard Neustadt compared the utility of historical analogies by scholars and policymakers to the justification for sex education classes: if teens will inevitably do the deed, why not help them do it better and more safely?

Not just better and more safely but more in general. The failure mode I’m most worried about is a singular analogy becoming the template for thinking about U.S-China relations. A diversity of analogies helps us get a better grasp of the situation. Falling in love is not just like falling in love. It’s also like soaring on the wings of eagles, like taking the first step on an icy lake, etc. For analyzing the current U.S.-Sino relationship, the addition of the Sino-Soviet split analogy is a helpful contribution.

Part III: The Clone Wars

I want to add two more historical analogies to the mix, both involving U.S.-Japan relations, which hopefully shed more light on how to grasp the U.S.-China relationship. The first is U.S.-Japan high-tech competition in the late 20th century. I’d wager that very few readers (and I definitely do not) remember the days when people thought we were in a Cold War with Japan! Here’s Samuel Huntington in 1993 for International Security (p. 75):

“Sophisticated analysts such as Joseph Nye argue that Japanese-American "economic competition can leave both sides better off, albeit not in every case and sometimes with considerable stresses." Hence it is "inappropriate" to speak of an "economic cold war" between the two countries. It takes only one side, however, to produce a cold war. Japanese strategy is a strategy of economic warfare. Japanese leaders regularly assert that economic competition is central to the relations among nations, that Japan must prevail in this competition, and that Japan is, indeed, prevailing in the competition.”

We see this theme often repeated in current discussion by prominent “China-watchers” like Bill Bishop and Bethany Allen-Ebrahimian, who runs Axios’s China newsletter. Regarding a May issue titled “The ‘new Cold War’ started in Beijing,” Matt DeButts noted that Axios China’s “anti-CCP tone is edging from ‘slant’ into straight-up ‘bias.” In the piece, Bethany makes many of the same talking points that Huntington made back in 1993. “A growing number of experts are warning against what they call a ‘new Cold War’ with China. But many Chinese Communist Party elites already view the rest of the world as a staging ground for competition between China and the United States,” she writes. In other words: They started it! This isn’t middle school. Imperfect analogies come with real costs.

It would be funny if it weren’t so sad and dangerous. As The Chicks croon in their recent album: “Gaslighter, big timer. Repeating all of the mistakes of your father.”

Part IV: Return of “Kill Japs”?

The second historical analogy I want to add into the mix is the U.S.-Japan relationship in the heat of World War II. We already analogize everything to war way too much. But the motivation here is to unpack how race and racism shaped the U.S. war in the. Pacific, compared to its approach toward the Nazis — with an eye toward warning against how race and racism affects current U.S.-China relations.

Much of this draws from John W. Dower’s 1986 book War Without Mercy, in which he lays out how the war in the Pacific Theatre was, in many respects, a race war. These are just brutal to read:

• The Japanese “were perceived as a race apart, even a species apart — and an overpoweringly monolithic one at that. There was no Japanese counterpart to the ‘good German’ in the popular consciousness of the Western allies (p. 8).”

• “The Navy representative to the first interdepartmental US government committee that was assigned to study how Japan should be treated after the war revealed himself to be a literal believer in Admiral Halsey’s motto ‘Kill Japs, kill Japs, kill more Japs.’ He called for ‘the almost total elimination of the Japanese as a race,’ on the grounds that this ‘was a question of which race was to survive, and white civilization was at stake.’” (footnote 61, p. 55)

• “There was a widely accepted belief that the Japanese, like ‘patient’ Orientals in general, thought in terms of millennia rather than centuries, and that this current conflict in the view of Japan’s leaders ‘was but a step in Japan’s hundred-year-war plan for world conquest.’” (p. 56) ****Hmmmmm….why does that sound so familiar? Oh maybe because it’s the conspiracy theory that Michael Pillsbury, who Trump thinks is “probably the leading authority on China,” and many others believe in — one that has been thoroughly torn apart.

Obviously, war escalates racial divides, and we are not at war with China. But racial divides can also escalate threat perceptions and the risk of war, and so many of the threads above can be seen in the Trumpian approach to China. When Trump says that almost every Chinese student that comes over to this country is a spy, or when Sen. Marco Rubio and FBI director Christopher Wray warn about China as a “whole-of-society threat,” it’s not that hard to draw a line to the “overpoweringly monolithic” perceptions of Japanese.

I don’t pretend to think that I can influence the Marco Rubios or Christopher Wrays of the world. Even less likely when it comes to the Peter Navarros, Steven Millers, and Steve Bannons. What I can try to do is appeal to the people who lend the sheen of legitimacy, the wisp of reasonability to their views — like, say, the people who write about espionage risks, the United Front, and tech transfer concerns.

Here’s what I’d like to say. There are no neutral takes. Your reports, ideas, Tweets — the “technology” — you are putting out into the world are value-laden, and the writers of takes, the gatekeepers of which takes get read must bear additional responsibility. To not just give the proper context or footnoted caveats — like how while there are some concerning cases of espionage, the VAST MAJORITY OF CHINESE STUDENTS IN THE UNITED STATES ARE JUST HERE TO MAKE A BETTER LIFE FOR THEMSELVES — but also to carefully consider how those with Trumpian values who are in power will contort and exploit these takes. Sometimes chasing the hype cycle and the clout is not worth it, and you can push ideas in more careful ways.

Regardless of what type of platform you have — from something as small as ChinAI to one as big as the NYT — we all have to be more careful of the analogies we feed and the ones we neglect. A diverse set of analogies is an important starting point on the road toward a more perfect understanding of the U.S.-China relationship. It’s not unlike falling in love.

Thank you for reading and engaging.

These are Jeff Ding's (sometimes) weekly translations of Chinese-language musings on AI and related topics. Jeff is a PhD candidate in International Relations at the University of Oxford and a researcher at the Center for the Governance of AI at Oxford’s Future of Humanity Institute.

Check out the archive of all past issues here & please subscribe here to support ChinAI under a Guardian/Wikipedia-style tipping model (everyone gets the same content but those who can pay for a subscription will support access for all).

Any suggestions or feedback? Let me know at chinainewsletter@gmail.com or on Twitter at @jjding99

Greetings from a land where translations are living breathing things…

Very grateful to all the people who engaged with last week’s translation on tech neutrality — shout out to Kohzy Koh and James Bradbury for some exhaustive fact-checking and editing. Many readers also took on the translation challenge for the really tough paragraphs on Chinese views of science & tech in the second half of the 19th century. Thanks to Yishu Mao, Brayden Mclean, and Ryan Soh for sharing their translations (p. 6 of last week’s translation).

…as always, the archive of all past issues is here and please please subscribe here to support ChinAI under a Guardian/Wikipedia-style tipping model (everyone gets the same content but those who can pay support access for all AND compensation for awesome ChinAI contributors).

Feature Translation: Lip Reading “The Bad Kids”

In the last few weeks a Chinese web drama about three children who accidentally film a murder, titled “The Bad Kids” (隐秘的角落), has caught fire. Here’s Sixth Tone on the rave reviews it has drawn:

“Crouching Tiger, Hidden Dragon” star Zhang Ziyi on Thursday praised the show as one of the first Chinese dramas she’s seen that rivals American and British productions in terms of quality. On IMDb-like review platform Douban, “The Bad Kids” has an impressive aggregate rating of 8.9 out of 10 based on over 440,000 reviews.

CONTEXT: This week, we translate a blog post from a fan of the series, Eury Chen, who dug into why there were certain lines from the episodes that had to be modified in order to pass China’s review process for TV dramas. They applied lip-reading AI to restore the original lines of dialogue:

“Two days ago, I watched the TV series "The Bad Kids" in one go, and the plot was quite exciting. The disadvantage is that in order for the series to pass the review (of the National Radio and Television Administration), the edited sequences for episodes 11 and 12 were disrupted, even to the point that lines were modified, so that there are several places in the film where the actor's mouth movements and lines are not matched, which makes the plot confusing to people. Therefore, I tried to restore the modified lines through artificial intelligence technology, thereby restoring some of the original plot, which contained a darker truth.”

KEY TAKEAWAYS:

• Eury’s lip-reading process:

1) Find places where the actors’ lip movements do not match up with the actual dialogue (mainly in the last episode)

2) Use the facemesh model (a package in TensorFlow) to obtain features of facial expressions (image below) to predict the the initial of a Chinese syllable (声母 -- usually a consonant like b in bao 抱) and the final of a Chinese syllable (韵母 -- i.e. the syllable other than the initial, so like -ao in bao 抱).

3) Try to find the best match between the choices of pinyin* and the context clues in the plot. It is too difficult to predict Chinese characters directly through the shape of the mouth.

*system of transcription for Mandarin Chinese sounds (e.g. bao for 抱)

• Why did the censors modify the dialogue?

Eury explains that the original dialogue violated two of the most important principles in the review process of Chinese dramas: 1. No unsettled cases; 2. Bad guys must be brought to justice.

An example of the original dialogue and post-censor modifications (the original Mandarin has GIFs that show the exchange). Note how the original dialogue conveys Zhu Chaoyang, one of the boys who witnessed the murder, in a much darker light:

MODIFIED Yan Liang: "Tell the police" 「告诉警察吧」

---> ORIGINAL = Yan Liang: “What now?” [那该怎么办]

MODIFIED Zhu Chaoyang: "As my dad hoped" 「像我爸希望的那样 」

---> ORIGINAL = Zhu Chaoyang: “Apart from having him* get caught...”「除非让他被抓...」*him refers to the substitute teacher they filmed murdering his elderly in-laws

MODIFIED Zhu Chaoyang: "Do you want to call the police?" 「你想报警么」

----> ORIGINAL = Zhu Chaoyang: “Don’t you want to get revenge?” 「你不想报仇么」 --- one of the pinyin outputs from the lip-reading algorithm: ni bu xiang bao chou me

• Broader implications:

As Jack Clark has written in previous issues of ImportAI, lip-reading has a broad range of uses, including assistance for the deaf or hard-of-hearing, advertising, but also surveillance. In October 2018, the Chinese Academy of Sciences and Huazhong University of Science and Technology, which the researchers claim is “currently the largest word-level lipreading dataset and the only public large-scale Mandarin lipreading dataset.” The TV theme returns in this case as well, since the dataset was built by annotating Chinese TV shows.

FULL TRANSLATION: Lip Reading the Bad Kids

Financial Transparency for Year 1 of Paid Subscriptions

About a year ago on July 7, I launched the paid subscriptions option for ChinAI. The promise was that subscriptions would go toward compensating ChinAI contributors for their translation work and expert analysis. In that post I also committed to donating 10% of any earnings to GiveWell. Thank you so much to all the subscribers — very cool to have 150+ folks financially invested in this community. You can join them by subscribing here.

Here’s the financial breakdown from this past year. Net revenue after Stripe and Substack took their cut was $6756.45 (screenshot below from my Stripe account):

For contributors and translators, I paid out $1550 to 13 people (some of whom generously donated their time). Taxes and other expenses brought me down to a net income of under $5000 for the year of paid subscriptions. So I donated 10% of that to GiveWell, which aims to direct donations to effective charities.

Anyways, just trying to shine a little bit of sunlight on how I’m taking care of my tiny little piece of land amidst the vast Internet. Thanks to everyone who’s helping it grow.

ChinAI (Four to Forward)

Must-read: 5 Takeaways from China’s Draft Data Security Law

Samm Sacks, Qiheng Chen, and Graham Webster digest five important developments in the draft law. It’s a really impressive piece that combines 1) distillation of a lot of complicated legal jargon, context from historical thorny issues in data security, 2) insights from key Chinese scholars and standards drafters working on these issues, 3) balanced and concise analysis, which avoids hype-chasing.

1. “The formation of a data classification system at the national level that would delineate different types of data for different treatment under various laws and regulations” — relates to forthcoming work on an “important data” standard.

2. Seeks to define a procedure “to specify how the state (e.g., the Ministry of State Security, the Ministry of Public Security, etc.) lawfully may access data from private sector platforms.”

3. Could be “the first national law that recognizes and even calls for the establishment of data transaction markets.”

4. Tries to address bureaucratic turf battles by clarifying jurisdiction on data protection

5. Tackles cross-border flows of data, including export controls on data and codifying reciprocity on “discriminatory” treatment of Chinese businesses.

Should-read: Balancing Standards — US and Chinese Strategies for Developing Technical Standards in AI

Unpacking a few thoughts on my recent commentary for the National Bureau of Asian Research on the importance of technical standards in US and Chinese AI development.

• First, the two NBR reports on standards that have really stood the test of time and are still essential overview of China’s approach to standards-setting are Kennedy et al. 2008 and Yao and Suttmeier 2004. You have to pay to read them I believe, but it’s well worth it.

• One of the things I wanted to work through was clarifying what is meant by “standards” — something that’s still a work in progress. I discuss 1) technical standards as int’l private standards (e.g. via bodies like the ISO/IEC); 2) int’l public standards (e.g. via intergovernmental organizations like the ITU — though these bodies are becoming more and more privatized); and 3) domestically-oriented government efforts to provide measurements, specifications, and benchmarks (which are also often discussed in the language of “standards). For example, NIST’s facial recognition vendor test helps establish certain benchmarks for evaluating algorithmic accuracy. Sometimes these become incorporated into an international technical standard down the line.

• Main point I want to get across that technical standardization involves a broad balance of different interests, beyond the narrow IP/competitiveness concerns that dominate existing discussions. In particular standards are important vehicles of international private regulation, and they will play an important role in AI governance.

Should-read: Thread on Douyin content monitoring process

Isabelle Niu provides a window into how Bytedance (owns Douyin & TikTok) regulates livestreaming content based on their detailed 2019 report. This process involves real-name registration + facial recognition (to prevent foreigners from borrowing Chinese IDs to register), coordination between AI systems and humans to do real-time monitoring of livestream content, etc.

Izzy 穆戈 @Izzy_Niu

THREAD: This seems really alarming so I did a little digging and found that apparently ByteDance, which owns Douyin & TikTok, explained how it self-regulates livestreaming content in a detailed 2019 report. I'll break down some of the main points of the report in this thread: 1/

Joshua Dummer @JoshuaDummer

Wife trying livestreaming on Douyin. Comes over to me in the middle of a stream and puts me on camera. After about a minute, notice pops up saying that foreigners are not allowed to appear on livestreams "without permission".

July 8th 2020

389 Retweets677 Likes

Should-read: Turbocharged China Chip Investments Spark Bubble Fears

For Reuters, Josh Horowitz and Samuel Shen give an informative update on soaring valuations ofr Chinese chip-related companies. Prior to reading this article, if you asked me about a promising Chinese AI chip company, Cambricon would have been one of the first companies I listed. Here’s an interesting nugget that caused me to update my priors: “Cambricon, an AI chip company set to list on the STAR market this year, in its prospectus said it booked revenue of 444 million yuan in 2019 and a net loss of 1.18 billion yuan. After losing business with Huawei, Cambricon now earns almost half of its revenue from a single municipal government branch.”

Thank you for reading and engaging.

These are Jeff Ding's (sometimes) weekly translations of Chinese-language musings on AI and related topics. Jeff is a PhD candidate in International Relations at the University of Oxford and a researcher at the Center for the Governance of AI at Oxford’s Future of Humanity Institute.

Check out the archive of all past issues here & please subscribe here to support ChinAI under a Guardian/Wikipedia-style tipping model (everyone gets the same content but those who can pay for a subscription will support access for all).

Any suggestions or feedback? Let me know at chinainewsletter@gmail.com or on Twitter at @jjding99