ChinAI #353: Year 8 of ChinAI
Eight things I learned about China's AI ecosystem over the past year of newsletter notes
Overview
Greetings from a world where…
ChinAI is now eight years old, which means it should know what day of the week it is, be able to tell the difference between right and left, and begin to take an interest in reading books (we were an early starter on that last one). *I want to thank the paying subscribers who have financially supported this work over the years, which ensures all ChinAI content is available to all.
…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).
Eight Takeaways from Year 8 of ChinAI
To celebrate eight years of the ChinAI, let’s review eight key takeaways from the past 50 issues of the newsletter. They roughly map onto my favorite issues from the past year.
The China Academy of Information and Communications Technology (CAICT) has solidified its place as “the Chinese equivalent of the AI Safety Institute.” Technically, as the linked DigiChina piece mentions, that title belongs to the China AI Safety and Development Association, which is a consortium that brings together CAICT and seven other groups. When I checked in on the organizations consistently updating their AI safety benchmarks (see ChinAI #315 below), however, CAICT stood out. They also regularly publish lengthy, in-depth AI governance reports (ChinAI #343).
The reality of China’s AI adoption: shallow, narrow, and slow. This shouldn’t be surprising for those familiar with the diffusion timelines of general-purpose technologies. Hopefully, by carefully tracing how DeepSeek was being disseminated (hint: it was not via software-as-a-service), this issue cut through some of the hype about China’s purported nationwide embrace of DeepSeek across hospitals and local governments. As I noted, “Since the frenzy of reports about fast and wide DeepSeek deployment, these all-in-one machines have not diffused past early adopters and attracted few repeat customers.”
There exists a massive and growing capital expenditures gap between Chinese and American tech giants. Stats from the issue: In 2020, this capital expenditure ratio of Chinese tech giant capex compared to US tech giant capex was 1:6, and by 2024, it had reached 1:10. This Jinduan Research Institute piece also called out Chinese tech giants for investing their profits in share buybacks and dividends, rather than capital expenditures.
If you read a lot of the initial coverage about China’s AI Plus Plan, for which implementation guidelines were released in August 2025, you could easily fall into two traps: 1) believing that the Chinese government had an enlightened approach to AI diffusion ; 2) equating the mere existence of a plan with actual progress toward that goal. This issue offered a more critical reading of the AI Plus initiative, based on this complex methodology I pioneered called “actually reading the text.” If you do carefully read plan, you’ll not only notice the lack of substance but also just have to laugh at certain sections (e.g., highlighting global cooperation as one of the six sectors for early AI adoption, or targeting 90% adoption rate of AI across the entire economy by 2030).
China’s draft companion AI regulations provide a window into the sometimes contentious relationship between the government and Chinese companies. This commentary by Chen He, a data compliance manager at a large Chinese securities firm, pointed out specific planks that AI companies would collectively push back on, including the requirement for “separate consent” for use of users’ interaction data in model training. The piece also referenced the impact of industry input into the final version of China’s generative AI services measures. In my analysis for this issue, I traced how the final draft of those measures watered down the mandate for AI developers to ensure the “veracity, accuracy, objectivity, and diversity” of their training data.
How have international industry associations contributed to China’s remarkable safety gains in certain high-risk technological domains, including civil aviation and nuclear power? The inspiration for this co-authored paper, published open-access in Review of International Political Economy, came from an interview with Sig Hecker, former director of Los Alamos National Laboratory in March of 2022. I had been talking with him about the history of U.S. sharing of permissive action link technology, and he recommended that I look at civil nuclear cooperation (not just nuclear weapons safety and security). For the next four years, I spent a lot of time trying to understand the role of international forces in China’s efforts to raise safety standards in high-risk domains.
Interestingly, I went in with the expectation that the main characters of the answer would be international public regulators like IAEA or ICAO; instead, the actors that drew my attention were international industry associations like WANO and IATA. This ChinAI issue summarizes the article’s key findings and spins it forward to implications for global AI governance.
Stop making blanket statements about how all Chinese people embrace AI. Based on a feature translation of an article published in The Mirror [镜相工作室], I shared the stories of various people who posted on Xiaohongshu under the hashtag #反ai (#Anti-AI), which had racked up 5.1 million views and 40,000 discussion threads as of January 2026.
Lastly, the post from the last year that I was most proud of writing:
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.
Details
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.
Any suggestions or feedback? Let me know at chinainewsletter@gmail.com or on Twitter at @jjding99
Source
Originally published at chinai.substack.com.
