@selinaxu_ Great paper, but the source materials are English-language websites of PD and Xinhua. The discussion of "domestic framings" seems to mean narratives about things happening in China, not for a domestic audience. Would be nice to see similar analysis of truly domestic narratives!
Tencent’s Hy4 system card is out and includes our independent pre-release safety evaluation. It’s the first system card from a Chinese frontier lab, and a meaningful step for third-party evaluation in China.
We tested loss-of-control and harmful-manipulation risks with four private in-house benchmarks. Compared with its preview, Hy4 shows overall safety improvements across both risk areas. We also found some complex agentic scenarios still leave room for further improvement, e.g., multi-agent resource competition.
Thanks to the Tencent Hy team for their trust and openness. We look forward to working with more frontier labs.
The most detailed model safety card I've ever seen from a Chinese tech company. Not your standard HarmBench either but discussion and details of frontier issues from open-weight safety to bio to loss of control.
Notably, Tencent says it worked with a little-known new AI start-up called Fangcun AI to test on their internal benchmarks for misalignment. Here's my profile of them and interview with founder Luke Li Xiaojian: https://t.co/AOtGDsVGCI
This is a great way into AI governance for anyone looking to get into the field. I learned a ton working alongside some seriously smart and driven people as a fellow over the summer.
Apply for GovAI's 2027 Summer Fellowship, Research and Applied tracks!
📍 Location: London, UK
📅 Deadline: 23:59 GMT Sunday, 22 November 2026
Fellows will join us in London to conduct a project on a topic of their choice, participate in an expert Q&A and seminar series aimed at developing a broad understanding of the field, and spend time forging connections with researchers and practitioners. Learn more below.
🔗 Learn more about the Research Track: https://t.co/f6lul6qfHi
🔗 Learn more about the Applied Track: https://t.co/ixSB0pG1sC
Right now 🇨🇳 AI policy has more open questions, live debates & bureaucratic tussling than any point since 2023.
I got into lots of that w/ @Gregory_C_Allen on his pod, one of the weedsier convos out there. Curious to hear what others are hearing on this.
https://t.co/58YvHHKOe2
🚨 New paper with Thao Pham!
How does Beijing frame AI for global audiences?
Analyzing 3,051 state-media articles (2014–24), we find a striking shift after ChatGPT’s launch: from technological achievements toward AI diplomacy and global rulemaking: https://t.co/UfYtPHZbCg
Here are some things you might want to know about China and AI:
- A "Chinese HuggingFace" might never become public
- Beijing’s criticism of Anthropic has been unusually direct, even by the standards of its past attacks on American companies
- Mythos, more than anything, changed how the CCP viewed AI
- Jiang Shigong is the most interesting non-dissident Chinese intellectual (admittedly, not on AI, but important to know nonetheless).
You can learn all of this, and so much more, in my latest interview with @JulianGewirtz.
Julian is a researcher at Columbia and served under Biden as the White House’s Senior Director for China and Taiwan Affairs and the State Department’s Deputy China Coordinator.
He’s also the author of the best book on the pivotal decade of modern China's history (the 80s), and blogs at https://t.co/OzMWQ7THBg
Julian was the person my relevant-expert friends told me to interview on China - and I'm glad they did!
Watch it wherever you get your podcasts.
Anthropic doesn't see itself as a "normal company." Dario Amodei believes both that 1) superintelligent AI could destroy humanity, and 2) our only chance at survival is to have his trillion-dollar tech company build it first.
This is not an intuitive way to think about either capitalism or social change, but it’s led Anthropic to the eve of one of the biggest IPOs in history. So it's more important than ever for the public to understand and scrutinize its idiosyncratic theory of change.
My debut for @NYTOpinion (gift link!): https://t.co/xsrh02LaBq
because it is infeasible to mandate that all small- and medium-size firms, spread across every sector of the economy, adopt artificial intelligence. Those companies would do so only if it made sense from a business perspective."
The "China winning diffusion" narrative is clearly overblown, but Jeff's argument seems to hinge on defining AI diffusion as the way the US does it, via cloud services and enterprise software, then conclude therefore that the US is doing better...
For example, I'd like to see way more evidence for this claim: "The centralized pressure that officials can use to effectively develop large-scale infrastructure projects is much less effective for general-purpose technologies like AI...
Greetings from Shenzhen, China, where Trump has just landed in a city imagined as a vision of liberal reform by Xi Jinping’s father — but has transformed into a reflection of Xi’s top-down control. Latest: https://t.co/mKIJ317lYv
Last week colleagues at Carnegie casually dropped the best+most detailed report on how to actually *implement* key ideas in frontier AI policy.
Think we need incident notification? Read the section on how to actually do it.
Congrats @Scott_R_Singer et al https://t.co/HsrsLrj1ee
@kristy_loke Sure. I'm talking about the proposals being put forward by China. To me, this isn't an article about wanting a deal with the US or slower AI development. It's fleshing out Xi's vision for global AI governance. People (not you) may see only the risk line and misread the signal.
While the language on risk is strikingly strong, the proposals look far away from a pause or a deal with the US. It supports wider access and using AI to govern AI. "There is no single standardized answer, and no single country or mechanism can accomplish it at one stroke."
Yesterday, the CCP's top theoretical journal (求是, Seeking Truth) published a major piece on AI governance and risks.
I think this is clearly the strongest & most urgent language on existential risks used in such an authoritative outlet. See below for comp to recent major doc.
@kristy_loke Not sure my observation contradicts your nice rundown here. I think it’s mildly pro pacing. But its key point is that the AI governance conversation needs to be broader than current US-centric proposals and should be respectful of different countries’ political systems, no?
Super cool (for China nerds) project from @kavocabo to track various tifa and their growing importance for policymaking in the Xi era https://t.co/Y8mT5waaht