1/A year ago, Chinese scientists published data from samples collected at a Wuhan market at the heart of the COVID outbreak - startling many, given it happened years ago. Then the data got retracted.

Maria Cheng and I tried to find out why. Our findings:
apnews.com/article/china-…

Can someone recommend a person or paper/book on China’s relationship with the UN?

Please feel free to recommend yourself or your work.

Breathtaking new artwork hanging in the Louvre this morning 🤌

Good piece on if/when China will pass a comprehensive AI Law, what that would look like, and the different players involved.
geotechnopolitic.substack.com/p/chinas-ai-la…

from Viola Zhou 's excellent piece restofworld.org/2024/tsmc-ariz…

Important note: this work of analysis by UN experts was greatly facilitated by the vital work of China Law Translate / Jeremy Daum (who referenced its translations) to map, translate, and navigate China's legal and regulatory landscape.

Go follow this important work.

Recommend reading Tony Peng latest newsletter on Llama-3’s likely impact on China’s LLM scene.

Largest impact likely to be on internal enterprise applications, with developers quickly building all kinds of applications.
open.substack.com/pub/recodechin…

Early returns for Llama 3's Chinese dialogue abilities aren't looking great.

Doesn't consistently answer Chinese questions in Chinese, even after agreeing to do it. Misreads / makes typos when repeating back characters to me. And most impt: doesn't seem to understand 肉夹馍!

A really moving piece of visual journalism that captures the tragedy in ways that change how you (re: I) relate to and process it:
nytimes.com/interactive/20…

Bit more on this law, which was a really interesting read! In Aug '23, a team of scholars from the Chinese Academy of Social Sciences (CASS) published a draft AI law. Stanford DigiChina Project covered this, w input from various experts. The CASS team have now released an update

see below for my take on the big-picture shift in 🇨🇳 AI governance over the past year, and how the state is trying to thread the needle / have its cake and eat it too.

AI policy scholars released an updated version of their 'expert draft proposal' for 🇨🇳's national AI Law.
Ch + Eng: zenodo.org/records/109741…

This is an update on the draft released last Aug & translated by Stanford DigiChina Project. More emphasis on exemptions for OS AI.

US vs China readouts: Assistant Secretary Kritenbrink’s Meetings in Beijing, the People’s Republic of China (PRC)

Interesting article on Chinese chip scalpers that came up through CSET's Emerging Technology Observatory tool this morning.

mycaijing.com/article/detail…

Couple of highlights 🧵

China continues to persecute family of dissidents unlawfully, finds report.'As recently as 28 December, Chinese authorities vowed to abolish the practice of collective punishment of the family members of convicted criminals.'
theguardian.com/world/2024/apr…

Now Jeffrey Ding has come through with a translation of this 'authoritative' AI safety benchmark:
chinai.substack.com/p/chinai-261-f…

The head of this AIIA committee led two earlier AIIA committees in 2019-2020 that fed into CAC's first two regulations on recommendation algorithms and deep synthesis. *Potentially* a preview of direction of future regs.

For more on how that works see: carnegieendowment.org/2024/02/27/tra…

New and 'authoritative' AI safety benchmark in 🇨🇳 this week. Put out by an AIIA committee linked to CAICT.

Covers content security, but also chem/bio risks and 'anti-human tendencies'(?). I'm sure it's crude but worth following (and someone translating).
mp.weixin.qq.com/s/3FcLBHCy_oVa…