1. Home
2. News
3. AI policy

A day after Amodei asked Washington to keep its AI lead over China, Xi offered BRICS an open source AI community.

China's five proposals for the bloc open with AI, and the one thing Xi asks the rest of the world for is a framework everybody writes together.

By Think Facility · September 13, 2026 · China AI policy Save

On September 13, 2026, Xi Jinping spoke at Session II of the 18th BRICS Summit in New Delhi and put five proposals in front of the other leaders. The first is about AI, and China wants to lead it.

Xinhua published the full text the same day. Read next to the argument the AI labs spent the weekend having, it points the other way.

Five initiatives, AI at the top

• Open source and inclusive AI
• Trade and investment facilitation
• Digital industry cooperation
• Intelligent manufacturing cooperation
• Science and technology talent development

Here's the first one, in his words: "China will be a pioneer in establishing a BRICS AI open source community, support the cooperation in developing and applying large language models, hold specialized AI seminars and training courses, and build an open ecosystem for AI."

Earlier in the speech (in a passage on global challenges) he asked countries to "develop AI for the positive and for good, and work faster to evolve a consensus-based global AI governance framework". He didn't name a body to write it. He didn't give a date.

Why the day before matters

On September 12, Dario Amodei published an essay asking AI companies to slow down. That plan only works, he argues, while American labs stay ahead of what he calls CCP-associated projects.

So the essay asking for restraint also asks Washington to "not sell powerful AI chips or semiconductor manufacturing equipment to China" and to "Crack down on unauthorized distillation by companies in authoritarian countries" (distillation is where a lagging company trains its own model on a frontier model's output). Xi's first initiative points the other way: a shared open source community, and cooperation on large language models across the bloc.

Where the two texts land in the same place

Amodei's third step is coordination with China, and he sorts possible agreements into four levels by difficulty. The second level is both sides testing their models before release for acute risks, which he says "could be done through a global standards body". He thinks "creating such a body is likely feasible".

Xi's phrase is a "consensus-based global AI governance framework". Amodei's is a standards body that tests models before release. Weirdly, neither man says who'd run his version.

What the speech leaves out

It names no rules. There's nothing on what an open source community would publish, and nothing on who'd check that anyone kept their word. Amodei's essay is blunt about that gap: an agreement "must either have ironclad verifiability, or must be limited enough that defection would not be militarily existential".

The harder part is what such a body would be allowed to look at. Amodei wants evaluators who can see a lab's training pipelines from the inside. Xi, in a passage about international rules in general (it isn't about AI), says those rules "must be written by all countries". I'm not sure how one body could do both, and neither text says.