ChinAI #55: A Reader's Inside Look at "Cold War 2" Meme Peddling

Plus, AI in China's Biopharmaceutical Industry

Welcome to the ChinAI Newsletter!

*** Thanks all for the feedback on the “tipping” model — I plan to lower the subscription fee to $12/month (still cheaper than Netflix’s standard monthly plan!) and then give it another week or two before transitioning over, so please continue to give feedback. Just to clear up any lingering confusion — I’m using The Guardian’s approach: everyone will still get every issue, but those who can pay for a subscription will help support access for all. This will enable me to also pay back contributors for great work like this week’s “Reflections” segment by Evan Jones.

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 the Cold War Meme by Evan Jones

*Evan is a PhD student in Government & Politics at the University of Maryland where he researches the different instruments China wields to influence other countries’ domestic and foreign policies. Check out his website here (I’d recommend his article on Chinese bureaucratic contestation of national roles in the South China Sea). In this week’s reflections segment, Evan distills his thoughts on the dangers of group-think and the Cold War meme. What follows is Evan’s text (lightly edited by myself):

Two weeks ago, I attended the Asia Policy Assembly (APA) in DC---a yearly gathering of Asia experts drawn from academia and the policy world. The discussion centers on current trends in Asian politics with the aim of informing current and future US strategy in the region. As one might imagine, Sino-US tensions dominated the discussion. This is not surprising. What did surprise me was the degree to which the “Cold War 2” (henceforth CW2) meme permeated all the plenary discussions, especially from some of the more prominent China-hands in the room.

After I returned from APA, Jeff sent out his newsletter and featured a debate section entitled: “Challenging the Gatekeepers.” His comments could not have been timelier nor resonated with me more. A few of his arguments were especially poignant given my recent experiences at APA and in other meetings around DC. Below I reflect on my time in DC in the hopes that it can further contextualize some of what has come up in this newsletter as of late.

While I agree there is good reason for China watchers and policy-makers to reevaluate America's long-standing posture towards the PRC, I disagree with those who seek to analogize about the current trajectory of Sino-US relations through a Cold War lens. The degree to which group-think has taken over the Beltway bubble shocked me. Maybe I am too green and naïve to DC’s modus operandi…But from my home base at the University of Maryland not far from downtown, I was under the false impression that this view was just one among a diverse chorus of voices.

Not so. That the US and China are locked in an existential battle appears to be the premise upon which all the discussions in the DC China watching community now stem. Longtime China hardliners and CW2 advocates take the growing bipartisan consensus on China as vindication and everyone else is forced to engage the issue on this turf.

Explicit employment of the CW2 meme varied at APA, but its underlying assumptions—China is an existential threat to the US; all of China's intentions, no matter their outward appearances, are aimed at undermining US supremacy; China has been at quietly at war with US since 1971, it’s time the US realize this and undertake [insert Cold War strategy here]; etc.—colored much of the rhetoric. Although these assumptions would have met stronger pushback a couple years ago, after Campbell and Ratner’s “The China Reckoning” piece they appear not only admissible, but taken as received wisdom.

Do the CW2 peddlers fit a pattern? Jeff previously emphasized demographics. Based on my perceptions at APA, the most distinct demographic delineation was generational. Older participants were more likely to invoke the CW2 frame and its associated strategies. Conversely, younger participants (whose presentations unfortunately are not in the link I provided) tended to be more nuanced and complex in their analyses. Of those whose presentations I attended or with whom I personally interacted, not one leaned upon anachronistic historical analogies nor implied that China posed an existential threat to the US.

I do not advance any conjectures as to why this is and instead leave it to the readers. Moreover, this is not to say there are not individual exceptions to this pattern nor that gatherings like APA are counterproductive. I encountered many people who did not fit this mold, and found the overall experience rewarding.

Why is this dangerous? Apart from the arguments Jeff laid out in previous installments, myriad cognitive biases plague the CW2 meme (and oftentimes historical counterfactuals). These include:

1.    Hostile attribution bias

2.    Anchoring

3.    Availability cascade

4.    Groupthink/bandwagon effect

5.    Confirmation bias

Take Made in China 2025 for example. The DC consensus primarily views this as a strategy to overtake US leadership in high-tech and advanced manufacturing (hostile attribution bias). While the PRC’s branding campaign in tandem with the “Chinese Dream” narrative certainly does little to reassure skeptics, the reality is the CCP must transition its economy in this direction if it wants to escape the middle-income trap and maintain its grip on power. Whether that’s possible is another debate, but, more importantly, the CCP’s intended audience is internal, not external.

The US can and should take issue with certain aspects of China’s strategy—namely forced transfer and championing of state-backed firms—for achieving its aims. But we must simultaneously acknowledge the CCP’s concerns and how our own biases foster further mistrust and suboptimal outcomes. Would we be nearly as critical if China were a liberal democracy?

Biases 2-5 form a pernicious package. As more and more China gatekeepers and people in the Beltway normalize the CW2 meme (bandwagon effect), the more prevalent it becomes in public discourse (availability cascade). As it snowballs, it primes policy-makers to think in zero-sum terms, focus on the most divisive features of the Sino-US relationship, and anchor their decisions around these features.

These biases manifest on the Chinese side too. Cadres and analysts witness the prominent trend in US discourse. They then attribute America with wanting to contain China’s rise [Bill Bishop has consistently hit on this without recognizing his role in exacerbating the CW2 meme]. Each side then takes the other’s rhetorical and policy response as proof of their respective beliefs (confirmation bias).

Why should we care? The stifling effect of the CW2 meme was tangible at APA. In terms of creativity, very few plenary sessions moved beyond discussing the key challenges in Sino-US relations. The most consistent policy recommendation was for the US to strengthen its alliances and partnerships in the region. While I do not necessarily disagree, this is an easy and—I would argue—vacuous recommendation absent details.

In terms of inclusiveness, one of my interlocutors mentioned a discussion he had with a young Chinese scholar who said she felt unwelcome and that the climate was hostile. Another person who had recently spent considerable time in China teaching and working said individuals from the intelligence community looked upon him with suspicion. While these are the impressions of only a handful of attendants with whom I interacted, both instances are very disconcerting.

The larger point is evident: The cold war meme fosters outmoded, zero-sum mentality and marginalizes the people who are best positioned to bridge the growing divide between our two countries.

The China watching community in America can seem small and insular. We can easily be tricked into thinking our primary audience is one another. Our impact is much broader, especially gatekeepers. My impression is that many in DC hold up Sinocism as the daily gospel on China. Bill may only be “reporting the news,” but his editorial decisions matters.

We must keep contesting the social hierarchies that shape our information ecosystem and those who are privileged with its management. More importantly, we must guard against the creeping biases that can permeate our analyses and public debate if we are not vigilant. The Cold War meme is a gross over-simplification. China and its relationship with America defy such reductionism.

Disclaimer: I will be on a funded fellowship from the National Bureau of Asian Research, one of APA’s sponsoring organizations, for 2019-20 academic year.

AI + China’s Biopharmaceutical Production Industry

This week’s first feature translation zooms in on AI applications in China’s biopharmaceutical industry, which is expected to grow from US$14 billion in 2013 to US$75.5 billion in 2022 with a compound growth rate of 20.6%. The goal is to get to “Pharma Industy 4.0” which involves automation of production equipment, informatization of pharmaceutical processes, intelligent production control. The most interesting sections of this mini-report were parts 4 and 5, which gave concrete cases of intelligent tech applications in biopharmaceutical factories:

Note Johnson Controls and GE as two international players involved in the biopharmaceutical chain. The translation also highlights an interesting case of Beigene (a Chinese NASDAQ-listed biotech company) purchasing GE’s KUBio factory, an innovative and ready-to-use modularized factory that can be designed and completed within 18 months in accordance with international Good Manufacturing Practices in pharma.

For a deeper look, READ FULL TRANSLATION: Future-facing Research: Current Status and Prospects of the Application of Intelligent Manufacturing Technology in the Biopharmaceutical Industry

Military AI + Maintaining Control of Command and Control

Re-upping this translation first featured in Issue #49it’s a post by a National University of Defense Technology researcher that connected the recent Chinese sci-fi blockbuster “Wandering Earth” to the need for early-warning systems of AI rebellions in military technology. Notably, the article called out the U.S. military’s loss of control of drones on the Iraqi battlefield. In the Google doc with the full translation I highlighted this line and commented that the author may be referencing the fact that 400 large U.S. military drones have crashed in major accidents around the world since 2001.

Christian Curridien, a defense analyst at RAND, commented on that thread (Note: these views are Christian’s and not that of his employer):

“More likely referring to several incidents in Iraq in which tracked ground drones reportedly turned their guns on their operators or other illegal targets. The stories were later debunked, but were widely circulated in China. Yet another example of how things are easily sensationalized, especially across language barriers. Anecdotally, it sounds like the systems may have suffered from unrelated mechanical difficulties, including wiring problems when going over rough terrain, though take that with a grain of salt as I haven't been able to find sources on that. Much larger armed Russian unmanned ground vehicles seem to have run into a variety of mechanical and technical problems that call into question their usefulness on the battlefield, totally unrelated to any loss of control.” Christian refers readers to these two Chinese articles on the issue as well as this one by Popular Mechanics.

READ FULL TRANSLATION: In “The Wandering Earth” Do We Encounter a Defection by MOSS? Artificial intelligence is advancing rapidly, intelligent command and control systems need early warning of rebellion

ChinAI Links

Shoutout to SupChina for updating their China Twitter 100 list to represent a greater diversity of voices — good example of a platform being reflexive and responsive to criticism.

Fantastic piece by Cate Cadell of Reuters on China’s data collection industry. Longtime readers of ChinAI may recognize similar themes in a ChinAI feature translation of a GQ China piece from six months ago, but Cate’s piece adds more context including good stats on the global market for ML-related data annotation.

Via Shen Lu’s recommendation, check out Karin Fischer’s “latitude(s)” newsletter, which looks at the global education landscape. This post on bearing witness to Tiananmen and Karin’s conversations with Chinese students is really great.

For more on China’s Medical AI market, see Caroline Meinhardt’s excellent analysis of some hidden challenges including openness to foreign competition, regulatory bottlenecks, and China’s underdeveloped and unstandardized electronic resource infrastructure for data sharing.

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 jeffrey.ding@magd.ox.ac.uk or on Twitter at @jjding99