ChinAI #292: The Misperception Spiral in US-China tech policy competition
Greetings from a world where…
if you know what these are, we can be friends and swap root canal stories
…As always, the searchable archive of all past issues is here. 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: Chinese Misperceptions of U.S. Technology Strategy
Context: While scanning through the ETO Scout tool, one article preview stood out to me: Chinese think tankers (two CAICT researchers) view America’s technology strategy as a “whole-of-government” approach centered on the executive branch’s list of ~20 critical and emerging technologies. During one of my meetings in Beijing this past week, a policy researcher affiliated with a big Chinese tech company also asked me about the role of the critical and emerging technology list (released by the White House in October 2020) in the U.S.’s technology strategy. It appears that this Communications World article (link to original Chinese) has gained a lot of traction.
Key Takeaways: These China Academy of Information and Communications Technology researchers construe the U.S. as having adopted a “high-intensity government intervention model” in technology strategy.
Yapeng Lu and Weiguo Wang write, “Over the past four years, the identification of critical core technologies has evolved from partial identification and deployment of single technologies to the identification of technology clusters and global strategic deployment. There have been rolling updates and the formation of three technology lists. The critical and emerging technologies list has become the ‘compass’ for the United States to build technological hegemony.”
To substantiate the U.S.’s “whole-of-government strategy” based on the critical and emerging technologist list, they point to the establishment of the U.S. State Department’s Special Envoy for Critical and Emerging Technology, the Department of Energy’s Office of Critical and Emerging Technologies, and the U.S. Citizenship and Immigration Services new “national interest waivers” that reference the critical and emerging technologies list.
Chinese think tankers perceive U.S. technology strategy to be much more coherent and coordinated than it actually is in practice. (Hey, why does that sound familiar?)
Let’s take a more detailed look at some of their examples. USCIS’s national interest waivers does make one reference to the critical and emerging technologies list (footnote 68), but the policy has a wide scope that covers general STEM fields (“furthering a critical and emerging technology or other STEM area important to U.S. competitiveness or national security”). The DOE’s Office of Critical and Emerging Technologies focuses on some of the CETs but certainly not all of them. In fact, the DOD has its own Militarily Critical Technologies Program, and a GAO report has found that these lists “rarely inform export control and other policy decisions.”
Funnily enough, I make the students in my GW class on AI and international politics do an exercise with the 2020 critical and emerging technologies list. They must come up with their own list of three technologies that are most critical to national security AND clearly specify their methodology. The second part is most important, and it’s missing from the U.S. process.
My first academic article “The Logic of Strategic Assets” argued that U.S. critical technology identification focused too much on an asset’s importance and neglected whether the technology was connected to an externality (firms or military organizations would not produce socially optimal results on their own).
For a much more comprehensive and rigorous approach to identifying critical technologies, I’m a big fan of the EU’s Key Enabling Technologies approach.
The CAICT researchers also emphasize that the U.S. employs a top-down approach to technical standards-setting in critical technologies. Again, this is a misreading. In my view, the U.S.’s advantage in technical standards development comes from its decentralized process. From my book (p. 200):
“Moreover, the US approach to AI standard-setting could prove more optimal…Market-mediated, decentralized standardization systems are particularly suited for advancing technological domains characterized by significant uncertainty about future trajectories, which clearly applies to AI…Government intervention therefore could lock in inferior AI standards compared with market-driven standardization efforts. In that light, China’s state-led approach to technical standards development could hinder the sustainable penetration of AI throughout its economy. For example, the Chinese central government plays a dominant role in China’s AI Industry Alliance, which has pushed to wrest leadership of standards setting in some AI applications away from industry-led standardization efforts.”
So, U.S. analysts tend to perceive that China’s tech policy is more coherent and whole-of-nation than it is in actuality. As this week’s article illustrates, Chinese thinkers do the same. Why should we care about this Misperception Spiral in US-China tech policy?
As the above passage illustrates, sometimes Doing Less is optimal. This spiral effect pushes both countries to do more — potentially leading to more counterproductive policies.
The CAICT researchers call for China to pay close attention and learn from the U.S.’s critical and emerging technologies list. However, one should be careful about just copy and pasting another country’s list, since it’s so difficult to predict which emerging technologies will be the most important. Just two short decades ago, the U.S. went all in on nanotechnology; now, it doesn’t even make a top 20 list of extremely broad technological domains.
One last note: It was interesting that this distillation of U.S. technology strategy didn’t mention technology risk management at all, even though it’s one of the three pillars outlined in that initial October 2020 list. In addition, the DOE’s Office of Critical and Emerging Technologist just established an initiative to spur U.S. leadership on safe, secure, and trustworthy AI systems.
FULL TRANSLATION: Understand the US strategy for critical and emerging technologies in one article
ChinAI Links (Four to Forward)
Must-read: Whack-a-Chip: The Futility of Hardware-Centric Export Controls
If you’re like me, then you only need the following sentence to be convinced to read this fascinating paper: “In this section, we reverse engineer candidate GPUs used by Tencent by analyzing representative code signatures in multiple Tencent codebases.” This was authored by Ritwik Gupta, Leah Walker, Andrew W. Reddie, who are UC Berkeley researchers (Risk and Security Lab; AI Research Lab).
Should-read: Chinese AI Companies Are Catching Up Despite U.S. Restrictions
Eliot Chen is doing some fantastic reporting for TheWireChina. His latest examines China’s progress in AI models, including a statistic that Alibaba’s Qwen was the most downloaded open source model this year. I learned a lot from comments by Adina Yakefu, a researcher at Hugging Face. Also includes some quotes from me.
Should-read: AI Safety and Automation Bias
A new CSET report analyzes automation bias as a crucial risk in deploying AI systems. Authored by Lauren Kahn, Emelia Probasco, and Ronnie Kinoshita, this report “explores automation bias and ways to mitigate it through three case studies: Tesla’s autopilot incidents, aviation incidents at Boeing and Airbus, and Army and Navy air defense incidents.” Having studied the Aegis and Patriot cases myself, I gained a lot of new insights from this report’s perspective.
Should-read: China Leadership Monitor Winter 2024 Issue
The latest issue of China Leadership monitor includes a piece by me on China’s AI Implementation Gap as well as an interview with me on Technology and the Rise of Great Powers.
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 an Assistant Professor of Political Science at George Washington University.
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).
Also! Listen to narrations of the ChinAI Newsletter in podcast format here.