Plus, Three Alliances Fight over Identity Authentication Standards + Brittleness of Face Scan Payments in China
|Jun 24||Public post|| 6|
Welcome to the ChinAI Newsletter!
*** First, a special announcement: I am considering adding paid subscriptions for ChinAI in a way that EVERY ISSUE still remains accessible for everyone, but subscribers will pay $16/month (no market research went into this so please give me advice on this) to essentially “tip” the work. Again, subscribers would not receive any extra benefits in forms of exclusive content (man, I’m really selling this well), but they would get that warm fuzzy feeling of supporting my work, the work of contributors who had been volunteering their time, and I plan to donate 10% of all proceeds to GiveWell, a nonprofit that uses evidence-backed analysis to identify high-impact giving opportunities with a “top charities” designation. Please give me feedback on this idea – for example, I’d be curious to know whether it would be relatively easy for people who work at companies to write this off as an expense. Two reasons I’m choosing this “tipping” model over conventional “pay-for-exclusive-access” model: 1) free newsletters maintain this ideal of the public commons where everyone has access to deliberate over the issues of the day, 2) since the whole point of the newsletter is to show the valuable perspectives coming out of China’ digital media scene, it would fit to adapt a digital tipping model for content creators that is very commonplace in China.
These are Jeff Ding's (sometimes) weekly translations of Chinese-language musings on AI and related topics. Link to subscribe here and archive of all past issues here. Jeff is a Rhodes Scholar at Oxford, PhD candidate in International Relations, Researcher at GovAI/Future of Humanity Institute, and Research Fellow at the Center for Security and Emerging Technology.
Reflections on Last Week
Make sure you’re strapped in and seated in a good posture position, because this one’s gonna be a doozy (3500+ words of analysis coming your way). Let’s handle some housecleaning first:
- In last week’s issue, I made the link between “one-in-three chance of nuclear war” and “humankind goes extinct” in 33% of reruns of history, andmany of you pointed out that nuclear war may not cause total human extinction. That’s an important distinction for those of us steeped in the existential risk literature but less so for those of us trying to make the point that repeating the Cold War is not an ideal situation.
- The second mea culpa is one for writing this flippant line, “When will the U.S. stop relying on old white men to shape its thinking toward China?” Many readers — interestingly enough, all of them old white men or those who will eventually be old white men — called me out for what they deemed “racist” language. This is an important enough topic that I’ll expand on it in my reflections and will also extend it to my debate segment challenging the current “gatekeepers” in China analysis.
With the benefit of hindsight, I would have been more precise in my choice of words and added more context to flesh out the two more important points I was getting at: 1) a need for more demographic diversity in among China “experts,” 2) the lack of this diversity has malignant consequences for US foreign policy.
In the paragraph preceding this line, I had picked out three people who shape(d) the White House’s China policy — all of whom white men, including one known for directing white supremacist beliefs into the mainstream. Let’s do a slightly more scientific scan of the people shaping Trump’s approach to China. Consider this proxy — a list of people seated around Trump’s desk to strategize before Trump’s first meeting with Xi Jinping at the G-20: Mike Pence, Steven Mnuchin, Wilbur Ross, Robert Lighthizer, John Kelly, Jared Kushner, Nick Ayers, Matt Pottinger, Peter Navarro, and Michael Pillsbury. Do you notice anything in common about any of those names?
There are costs to this mono-representation. Maybe if there was a Chinese American in rooms like these, they could convince Trump to not go around saying that “almost every” Chinese student who comes to the U.S. is a spy. Maybe that Chinese American could share with other advisers about how his own parents came to the U.S. as international students from China, confronting them with the humanity and individuality of each student. Perhaps, if there were more demographic diversity in Trump’s administration as a whole, the U.S. State Department’s policy planning staff would not be drafting memos on grand strategy about how the coming conflict with China is “the first time that we will have a great power competitor that is not Caucasian.” Just a thought: maybe if there was someone that wasn’t white in the room it would be harder to see the U.S. as a Caucasian great power? I don’t know, I’m just spit-balling here. And also, maybe, just maybe, having more diverse faces in the room would dampen discrimination against Chinese and other Asian-Americans who are disproportionately charged under the Economic Espionage Act, receive much longer sentences, and are significantly more likely to be innocent than defendants of other races.
Look, I definitely appreciate and take the criticism of the readers that my flippant comment was essentializing (like all of us, old white men contain multitudes and have their own struggles). For instance, Iain Johnston has a very different and more evidence-backed view of Chinese foreign policy and strategy. I get how my comment came across as a cheap shot against all white men who are in this space, and that’s not my intention. I will challenge a few specific white men in the debate section below but I hope to do so in a more kind and nuanced way.
In one of the stories in the The Broken Stars anthology edited by Ken Liu, a must-read ChinAI link from last week, there’s an English translation of a Pushkin poem:
Life's deceit may fortune's fawning // Turn to scorn, yet, as you grieve // Do not anger, but believe // In tomorrow's merry dawning. // When your heart is rid at last // Of regret, despair, and fear, // In the future, what has passed // Shall in kinder light appear.
For the future I’m going to try to make my rebuttals “in kinder light appear.”
Debate Segment – Challenging the Gatekeepers
Kindness does not mean the absence of confrontation and debate. Before I start going after folks, for the record, when I publicly rebut the arguments of others, my intent is not to "name and shame" or brag about how my perspective better. I only find it worthwhile to call out those who: a) I respect and think will engage in fruitful deliberation, b) have outsized influence on issues that matter to me. Last week’s rebuttal was important in that respect because many of us who study the governance of technology and AI are deeply worried about memes like the "US-China Tech Cold War" because they position tech development as one tied to civilizational supremacy, incentivizing powerful states to cut corners on things like AI safety.
This week’s argument is two-fold: 1) The need for more demographic diversity also extends to the “gatekeepers” of analysis on China (those who frame and edit the conversation about China), 2) There should be more push back against and competitive alternatives to Bill Bishop’s Sinocism newsletter because it is a platform that has hardened ideologically in its framing of the conversation on China in a way that is counterproductive for the U.S. (and Western) approach toward China. This is exemplified by his role in pushing the U.S.-China Tech Cold War meme.
On the first point, it’s not just Trump’s broken cabinet that needs repair in terms of demographic representation. Take this list of “100 Accounts You Should Follow” for China Twitter. As Jin Dong, who does comms for @CNStoryTellers, points out, the list was mostly white and mostly male (68 men, 36 women). This despite the fact that SupChina wrote a disclaimer at the top of the list saying that it tried really hard to balance the list to highlight the diversity of voices. Sadly, this leads one to conclude that if this were a list of the “100 most influential voices” or the “100 most important gatekeepers” on China Twitter (without any intentional balancing for diversity), it would be even more skewed.
Editorial positions are especially important in framing which voices are heard on China. I did a quick and dirty scan of those who held editorial leadership positions in this list of 100 and 8/10 were men. Somehow, more women are writing about and doing interesting work on China (e.g. the two female editors were Karoline Kan who runs Chinadialogue and Manya Koetse who runs What’s on Weibo) but their voices are still being drowned out by the overly-confident, under-prepared male “expert.”
This is not acceptable. As shown by NüVoices, a directory of nearly 500 female experts on Greater China, this also does not need to be the case. For every week’s four ChinAI links, I try to feature female voices in at least half. I don’t advertise this because this isn’t about virtue-signaling; rather it’s about being conscientious about which voices you are elevating and the importance of having a diversity of voices from different perspectives. I’m no saint on this, as evidenced by my apology in the beginning of this ChinAI issue, and we all have to work to be better.
I stress the role of editors because progress is not made just by letting more diverse voices through the gate; it’s made sustainable by having more diverse gatekeepers. Let me repeat that for the people in the back or those who have their earbuds in: It’s not just about letting more diverse voices through the gate, it’s about having more diverse gatekeepers.
Speaking of white male gatekeepers, this week I engaged with three of them on the question of the US China Tech Cold War meme — including Bill Bishop, the so-called the “China-watcher’s China watcher” and creator of Sinocism (“the presidential daily brief for China hands).
I’m not going to get into the weeds of that debate, as I already made the case last week for why the New Tech Cold War Warriors rely on descriptively bad assumptions and promote prescriptively bad narratives. I will note how baffling I find it that the leading China tech journalist for the New York Times (Paul Mozur) and one of the leading analysts on China tech who gets quoted in almost every other article (Paul Triolo) are almost giddy about “being right” and “starting” the US-China Tech Cold War arms race meme. Without any sense of absurdity or irony, Triolo refers to Mozur as “the Wozniak” of the meme and himself as taking on “the Jobs role.”
The bigger fish to fry here is Bill. Not to take the analogy too far but if Triolo and Mozur are the Jobs and Wozniak of the US China Tech Cold War meme, Bishop is the John Sculley III (the expert marketer that grew Apple’s product lines). Joining Triolo and Mozur on the side of the “don’t worry, there’s nothing historically inaccurate or dangerous about the meme” camp, Bill commented in response to my thread, “But what if CCP-led China already thinks it is in a civilizational conflict? America is not the aggressor, where is the criticism of CCP-led China’s responsibility for the current state of the US-China relationship? A journalist & a researcher are not the reason we are where we r.”
Perhaps I am overstating the influence of a journalist and researcher, but Bill may be underestimating his own influence in promoting this US-China Cold War meme. I already pointed out last week that he ended his last Axios China with this headline as one of his essential eight points: “A New Type of Cold War.” I no longer subscribe to Sinocism after it switched to paid subscriptions, but from my read of the free Axios issues saved in my inbox, Bill has consistently pushed this Cold War narrative even when it doesn’t fit at all.
Three additional examples. Exhibit A: In an April 6, 2018 issue about China’s new defense minister visiting Moscow, Bill somehow in his thought bubble connects this incident to the U.S. being in “a new Cold War with China.” He writes “Now that the U.S. is descending into a new Cold War with Russia, and embarking on what may end up as a new Cold War with China as well, maybe history has not ended?” I didn’t even notice this until now, but I guess Bill is also operating under the assumption that we are in a new Cold War with Russia too? Who does Bill think we are not in a new Cold War with?
Exhibit B: In a post titled “The signs of a new Cold War,” Bill highlights one big thing, “China foresees new Cold War with U.S.” What he is actually referencing is a quote from Wang Yong, director of the Centre for International Political Economy at Peking University, who said, “So now China is preparing for a prolonged battle – and also for the worst-case scenario of a new cold war, or even a hot war.” First of all, Yong is not China, and China is not a monolith. Second, making preparations for the “worst-case scenario” of a new Cold War does not mean one foresees said scenario occurring, and it definitely does not mean that “worst-case scenario” describes what is happening now.
Exhibit C: Bill recently got very excited about China Law Blog’s interpretation of an article from Qiushi (the CCP’s official ideological mouthpiece) on the U.S.-China trade deficit, and shared it on Twitter by reiterating the headline “Does China WANT a Second Decoupling? The Chinese Texts Say That it Does” and commenting “that recent qiushi essay was disturbing.” Wang Feng, editor of FT Chinese (and one of the few non-white people on that list of ten editors I mentioned above), pushed back on Bill and China Law Blog’s read of the Qiushi essay. His interpretation of the central point, which I share, was that “decoupling is wrong and the US should not push China to the point of decoupling,” not that China wants a second decoupling. Also, again Bill takes one Qiushi article and turns it into “The Chinese texts” (which is plural for some reason) say China wants a second decoupling.
Here’s my challenge to Bill: defend your marketing of the US China Tech Cold War meme. Read my argument in ChinAI issue #53 which lays out the case for why journalists and analysts and gatekeepers should not use the meme (it is both descriptively inaccurate and prescriptively dangerous), which I backed up with 20+ sources, one of which was my 4300+ word written testimony before the leading Congressional commission on US China relations. If you do, I’m happy to feature it on ChinAI or continue the debate in another medium.
I’m being hard on Bill here because he plays an outsized role on shaping the assumptions behind not only the current people working on China but also the next generation. During my undergraduate studies at the University of Iowa, where my interest in Chinese politics and foreign policy really blossomed, I “grew up" on devouring issue after issue of Sinocism and listening to Sinica with Bill, Kaiser, Jeremy, and David. Bill’s Sinocism is the go-to-source for anyone (ranging from the leading expert in a subfield of Chinese foreign policy to the political risk analyst for a Fortune 500 company) looking to get a comprehensive overview of China analysis — it’s a well-deserved position for someone who obviously puts in the work and has the years of experience to back everything up. However, I do think it's important to have pushback and a diversity of viewpoints especially on issues that are worthy of debate (e.g. whether the Cold War meme is a useful and/or accurate analogue for the current US-China relationship).
For that reason, I’ve been surprised by the lack of pushback by China scholars and experts against Bill’s recent ideological hardening — a hardening that is evidenced by those three exhibits but not limited to them (this is outside the scope of this post and something I will need to research more — in fact, a good project would be to do NLP analysis of Bill’s Sinocism issues). My half-baked hypothesis is that many people are hesitant to directly rebut Bill is because they are afraid of sacrificing the chance to get their work plugged in Sinocism, one of the most important gates to a wider audience for those doing research and analysis on China. My report on “Deciphering China’s AI Dream” probably doesn’t get as much traction as it did if it didn’t appear as one of Bill’s essential eight in Sinocism. It’s a weird phenomenon to have such a singular gate in a field. There’s no single gate for American studies or analysis of Germany or Japan.
Again I’m not sure about my hypothesis in this section, as it’s a half-baked idea I’m throwing out there. What I am absolutely sure about is that the “China-watching” community needs to have more gates and more diverse gatekeepers. Here’s my humble proposition that ChinAI is emerging as one of those gates, and I plan to redefine what it means to be a gatekeeper.
Always good with people skipping past the debate/rant sections to get to the substance, and this week’s substance is meaty (18 page, 6000+ word investigative report on the state of face scan payments in China). The intro anecdote that hooks the reader: someone’s roommate scanned their face while they were sleeping and transferred all the money out of their mobile payment system, revealing the vulnerabilities of the “face scan payment” functions of Alipay and WeChat pay. We translated the first half in Issue #49 which delves into the details about this specific case and what caused it. The now-completed second half reveals some fascinating insights into how “face scan payments” interact with a broader conflict among 3 strategic alliances (one led by Alibaba, one led by Tencent, and one led by an international mix).
Three takeaways I want to emphasize: 1) the security risks of face scan payments, 2) the importance of standards alliances, 3) this article being another example of strong investigative-style/advocacy-style reporting in the China tech space.
1. Security Risks of face scan payments
As a security measure, face scan payments should be able to detect whether the individual’s eyes are closed or not, to prevent theft by way of scanning one’s roommate’s face while they are sleeping. Jiqizhixin randomly collected and tested the twelve flagship smartphones of thirteen individuals to see whether WeChat and Alipay could be accessed by scanning users’ faces with their eyes closed.
According to the test results in the above figure, in the closed-eye state test for Alipay's “face scan payment” function, more than 33% (4 units) of the mobile phones can pass the payment in the closed-eye state. The writer blames Alipay’s choice of a 2D face recognition scheme (rather than a 3D one used by WeChat) which can’t achieve accurate live-body detection.
Only four of the phone models supported WeChat’s face scan payment, and 75% of them also allowed payment to process in the closed-eye state. Funnily enough, after contacting WeChat about this, the writer and colleagues repeated the tests with the phone models and now all of them prevented “face scan payments” with closed eyes.
2. This is a fight among three powerful strategic alliances over technical standards in identity authentication. As the writer states, “the focus area of conflict and competition…is the standardization of the entire process of financial payment.” There are three standards alliances and the one who gets the standard will also get the dominant market position: 1) FIDO Alliance, established in July 2012, supported by the Bank of China, ICBC, China CITIC Bank, Mobile’s cmpay, and a range of international giants (e.g. Microsoft, Google, Samsung, etc.); 2) the IFFA Alliance, started by Ant Financial in 2015 after Alipay left FIDO, which includes the China Academy of Information Communications Technology; 3) SOTER – an open source standard announced by Tencent in August 2017
3. The article’s author, whose internet name is 四月(April), is refreshingly combative and willingly to really go after these companies.
While she admits that companies may have some reason to tradeoff security for convenience as users think that if you need to do an additional test to prove that you are a live body, then the "face scan unlock" aspect is too cumbersome and delays the authentication time too much. She concludes with a strong statement that “it is currently the consumer and the consumer that bear the hidden dangers and security risks of the mobile phone face scan payment.” In response to Alipay’s comments that their facial scan is equipped with “gaze recognition” which effectively prevents access if a subject’s eyes are closed, she writes, “This response and slogan are especially pale and feeble at this moment.”
I had the pleasure of working with Shen Lu at the University of Iowa on campus initiatives to help support international students. Her piece for the NYT on the huge gap between the CSSA that Chinese students know and the CSSA that Mike Pence and others describe is a must-read.
DigiChina is putting in some seriously awesome work to correct the language asymmetry in AI policy research, doing a one-day turnaround on two sets of new AI principles, one released by an AI industry alliance group and another by an expert committee, see their June digest.
Some have called it the “Chinese version of ‘This American Life.’” Brush up on your Mandarin by listening to the 故事FM podcast. Also, a fun project for some aspiring subscriber: automatic transcription and translation of Chinese-English podcasts. Why is that not a thing?
Cool article in Global Policy by Maaike Verbruggen on how advances in civilian innovation do not automatically translate into major advances or rapid diffusion of Lethal Autonomou Weapon Systems.
Thank you for reading and engaging.
Shout out to everyone who is commenting on the translations - idea is to build up a community of people interested in this stuff. You can contact me at email@example.com or on Twitter at @jjding99