AnswersPublished12 sources
The U.S. and China need a common way to measure AI progress
The U.S. and China should build a shared, regularly updated picture of what AI can do and how quickly it is changing—not necessarily agree on identical rules.
Published byEdited with GPT-6 LunaImages generated with GPT Image 2
Research answer
Key takeaways
- The U.S. and China should build a shared, regularly updated picture of what AI can do and how quickly it is changing—not necessarily agree on identical rules.
- A common measurement framework could give both governments a clearer basis for discussing risks, even when their policy priorities differ.
- A report that more than 90% of new Chinese microdramas in the first eight months of 2026 were AI generated illustrates why governments may have less time to prepare than they assume.[7]
AI Prompt
openai.comCreate a landscape editorial hero image for this Studio Global article: Why should the United States and China establish a shared way to measure the rapid improvement of AI capabilities before deciding how to gov. Article summary: The central argument is that Washington and Beijing need a shared understanding of how quickly AI capabilities are improving before they can make sound decisions about governing them. The anticipated Shenzhen meeting off. Topic tags: general, education, news, general web, user generated. Style: premium digital editorial illustration, source-backed research mood, clean composition, high detail, modern web publication hero. Use reference image context only for broad subject, composition, and topical grounding; do not copy the exact image. Avoid: logos, brand marks, copyrighted characters, real person likenesses, fake screenshots, UI text, readable text, watermark
AI governance is harder when governments lack a shared picture of what the technology can do and how quickly that is changing. The United States and China do not need to agree on every rule to compare evidence about AI capabilities. Establishing that common ground could make their discussions more practical—and help them respond as AI affects more kinds of work.7
Measure capabilities before debating rules
A useful first step would be a shared, regularly updated way to describe changes in AI capabilities. As a practical proposal, that could track what systems can reliably do, how their performance changes over time, and what it takes to deploy them. These are possible measures to discuss, not a framework the two governments have already agreed to.
Shared measurements would not settle questions about acceptable uses, regulation, or national priorities. They could, however, help both sides distinguish disagreements about the evidence from disagreements about policy. The case for doing this soon is straightforward: the cited opinion article argues that AI is advancing quickly enough that governments may have less preparation time than they expect.7
Shenzhen talks offer a chance to keep the dialogue going
A meeting between Presidents Xi Jinping and Donald Trump is expected in Shenzhen on November 18–19, 2026. Before it, technical discussions on AI are scheduled for November, according to the Brookings event listing.1 The listing also says U.S. Treasury Secretary Scott Bessent and Chinese Vice Premier He Lifeng agreed to establish a formal AI dialogue, including an incident line for communicating about AI safety incidents.
1
That gives the countries a channel for continued engagement, but a summit alone cannot keep a shared understanding current. Ongoing technical discussions could help maintain a common evidence base between leaders’ meetings. This is a case for practical cooperation on measurement, not a claim that either side has accepted a particular set of AI rules.
AI-generated microdramas show why the stakes are broader than AI labs
The South China Morning Post opinion article reports that more than 90% of new microdramas in China during the first eight months of 2026 were AI-generated, without cameras or live actors.7 The figure is a reported claim from that article; the available source material does not independently verify it.
Even with that caveat, the example points to a wider question for policymakers: AI’s effects may reach creative work and production processes once considered difficult to replace. Measuring capability changes can help governments notice such shifts earlier. It cannot by itself determine how to support affected workers or set protections, but it can give those policy discussions a firmer starting point.
What a practical first step could look like
The near-term goal need not be a comprehensive agreement on AI governance. Washington and Beijing could use technical dialogue to identify a small set of capability measures, clarify how those measures are assessed, and revisit them as systems change. They could then use that shared evidence to inform separate debates about safety, economic effects, and appropriate rules.
The principle is simple: measure together where possible, even when policy choices differ. With a Shenzhen meeting expected soon and a formal dialogue already described by Brookings, the opportunity is to turn continued contact into a more consistent way to understand AI’s pace and reach.1
Studio Global AI
Continue your research
This page includes a source-backed answer you can continue inside Studio Global.
People also ask
What is the short answer to "The U.S. and China need a common way to measure AI progress"?
The U.S. and China should build a shared, regularly updated picture of what AI can do and how quickly it is changing—not necessarily agree on identical rules.
What are the key points to validate first?
The U.S. and China should build a shared, regularly updated picture of what AI can do and how quickly it is changing—not necessarily agree on identical rules. A common measurement framework could give both governments a clearer basis for discussing risks, even when their policy priorities differ.
What should I do next in practice?
A report that more than 90% of new Chinese microdramas in the first eight months of 2026 were AI generated illustrates why governments may have less time to prepare than they assume.[7]