ChinAI #229: Special Topic AI Security

Alibaba and CAICT researchers analyze domestic & int'l approaches to AI governance in terms of specific policies and standards

Greetings from a world where…

eating shengjian bao for breakfast at grandma’s is a luxurious life

…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: Alibaba and CAICT researchers analyze domestic & int'l approaches to AI governance in terms of specific policies and standards

Context: The winner of last week’s Around the Horn vote: an article from the China Information Security, an influential journal in the field of information security. Co-authored by researchers from Alibaba’s standardization department and the China Academy of Information and Communication Technology (a government think tank), this article analyzes domestic and global approaches to AI governance, organized around specific policies and technical standards.

It’s funny, when you look at the number of views racked up by all 10 options on WeChat, this article was dead last. At this point, I shouldn’t be surprised when ChinAI readers end up picking the most “boring” option. It’s fun to find your people.

The article covers four main topics: their perspective on global AI governance policies, how China’s AI policies tackle governance and security issues, important domestic and international standards related to AI governance, and recommendations for China's AI governance policies and standards.

A Key Takeaway for each topic: 1) Perspectives of CAICT and Alibaba researchers on global AI governance policies

• They emphasize that both the EU and US employ a collaborative governance model that has transitioned more toward “government + market”-led: “The latest development trend shows that enterprises are gradually shifting from passive acceptance of governance rules to active participation in the formulation of rules.” The article describes specific U.S. measures as avoiding “one-size-fits-all, excessive intervention.”

2) China’s AI policies and governance/security issues:

• This section collected some key documents I was familiar with, such as the latest draft generative AI regulations, as well as others I need to look at more closely. The latter group includes the 2021 Opinion on Strengthening the Ethics and Governance in Science and Technology, the draft “Trial Measures for Ethical Review of Science and Technology” (just released in April 2023), and the “Ethical Norms for New Generation Artificial Intelligence.”

3) Key domestic and international standards related to AI governance

• The authors give details on different working groups in a key international standards committee (ISO/IEC SC42). For instance, in the trustworthiness working group, Germany lead work on the AI risk management standard, the UK leads work on a survey of topics related to trustworthiness, and China leads work on standards for the controllability of automated systems (ISO/IEC TS 8200).

4) Recommendations

• From my read, the article seemed to advocate for China to be more like the U.S. with respect to AI governance policies and standards. The last section recommended a “transformation from the ‘hard law’ model dominated by vertical control to the multi-party collaborative governance model that integrates industry self-discipline with a combination of ‘soft law’ and ‘hard law.’”

***I made a lot more annotations than usual in the full translation, because I think this article is a very useful resource for summarizing key policies and standards in AI governance, to date. In the comments of the Google doc, I’ve also included links to English translations of Chinese-language documents referenced by the CAICT and Alibaba researchers. In the section on important Chinese technical standards for AI governance, I noted that the earliest published document highlighted by the researchers is the White Paper on AI Standardization (2018 Edition). Funnily enough, I translated this white paper for the very first issue of ChinAI.

FULL TRANSLATION: Special Topic AI Security | Analysis of Domestic and Foreign Policies and Standards on AI Governance

ChinAI Links (Four to Forward)

Should-apply: GovAI winter fellowships

My old stomping grounds are recruiting for winter fellowships, which provide an opportunity for early-career individuals to spend three months working on an AI governance research project, learning about the field, and making connections with other researchers and practitioners. They are looking for candidates who are strongly considering using their careers to study or shape the long-term implications of AI. Applications close on July 23rd.

Should-read: Washington is Suffering from Policy Capture, Not Groupthink, on China

In The National Interest, David M. McCourt argues:

This shift in predominant views of China—from potential partner to rival, and increasingly toward the de facto enemy—is better described as an exercise in “policy capture” by anti-engagers. Those arguing for a reinvigoration of engagement and the careful management of America’s competition with China are, at most, politely tolerated, and are certainly not in positions of authority. The individuals who are in power have concluded Washington and Beijing are locked in a decades-long ideological conflict.

Must-read: Mother Sauce

This longform book review by Marian Bull, about the act of cooking a recipe as an act of translation, made me think the most this past week.

Should-read: How modern thinkers are expanding their ideas on Substack

It was really fun to peruse this roundup of academics on Substack. Some very cool stuff happening at the intersection of academia and blogging.

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).

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