--- title: "China's AI Cooperation Action Plan names eight steps to bridge the digital divide" date: 2026-11-01 category: Policy & Governance site: NeuroAI canonical: https://neuroai.site/a/na-policy-ai-cooperation-plan language: en --- # China's AI Cooperation Action Plan names eight steps to bridge the digital divide > Released at WAIC 2026 in Shanghai on 17 July, the NDRC-led AI Cooperation and Development Action Plan sets out eight concrete actions for international AI collaboration, plus a 10-story case collection. A hospital in a country with no domestic AI lab. A classroom where the textbook is a decade out of date. A farmer who cannot read the weather model that decides his harvest. China's newest AI foreign-policy document is, on paper at least, written for them. ## Key takeaways - Released at the World AI Conference (WAIC) in Shanghai on 17 July 2026 by the National Development and Reform Commission together with other departments. - The "AI Cooperation and Development Action Plan" (《人工智能合作发展行动计划》) lists eight concrete actions: quality data supply, inclusive compute, open-source sharing, deep empowerment, talent cultivation, joint rules and standards, security governance, and "AI for good." - It explicitly responds to United Nations calls for AI international cooperation, bridging the digital divide, and AI for sustainable development. - A companion case collection, "China's Wisdom, Benefiting the World (2026)" (中国智·惠世界(2026)), was released the same day with 10 cooperation stories — the third year running. - The plan is a policy statement of direction, not a binding treaty or a program with disclosed funding amounts. ## The eight actions, plainly The plan organizes its cooperation agenda into eight named actions: - **Quality data supply** — making useful training data more available across borders. - **Inclusive compute** (智能算力普惠) — extending affordable computing power to partners who lack it. - **Open-source ecosystem sharing** — opening models and tools rather than keeping them behind walls. - **Deep empowerment** — applying AI in health, agriculture, industry, and public services abroad. - **Joint talent cultivation** (数智人才共育) — training and exchange programs. - **Joint rules and standards** (规则标准共建) — building common technical and governance frameworks. - **Security governance collaboration** — coordinating on safety and risk. - **"AI for good"** (人工智能向善) — framing the whole effort around human-centered, beneficial use. The throughline is that AI should be treated as a shared global public good, with practical mechanisms (data, compute, open source, talent, standards) rather than only high-level principle. ## Why the "Global South" framing matters Bridging the digital divide is the plan's stated emotional and political core. Most of the world's governments do not have the compute budgets, talent pools, or model infrastructure of the United States or China. By packaging cooperation around open source and inclusive compute, the document positions China as a supplier of accessible AI infrastructure to countries that would otherwise be buyers-only. That framing sits in deliberate contrast to two other postures in the global AI race. One is the closed, paywalled model ecosystem, where the best systems are gated behind commercial licenses. The other is the export-control approach, where advanced chips and models are restricted rather than shared. The plan's bet is that offering open weights, shared datasets, and affordable compute is a more durable way to set technical and governance defaults in emerging markets — exactly the kind of soft-standard-setting that shapes whose tools become the regional backbone for a decade. Whether that translates into deployed systems depends on follow-through that the plan itself does not specify. It names intentions, not contracts, and the hard part — who pays for the compute, and on what terms — is left to later agreements. ## Reading the signal For all its diplomatic language, the plan is best read as an operational sequel to China's earlier, higher-level calls for global AI governance. Where those set principles, this one lists mechanisms. The eight actions double as a menu of what Beijing is willing to offer, and a checklist of what partner countries can ask for. The real test is whether "inclusive compute" and "open-source sharing" become funded, repeatable programs or remain aspirational headings. ## The case collection Alongside the plan, the NDRC released "China's Wisdom, Benefiting the World (2026)" (中国智·惠世界(2026)), a collection of 10 stories about China's international AI cooperation. The publisher noted this is the third consecutive year the collection has appeared, and that this edition covers a wider range of countries and a richer set of "AI plus" application scenarios than prior years, with a stronger emphasis on the "people-centered, AI for good" theme. The collection is promotional in nature — a showcase, not an audit. Treat it as illustrative evidence of where Beijing says its AI diplomacy is active, not as a verified ledger of outcomes. ## Honest limitations This article draws on the official government news release carried on the State Council portal (sourced to Xinhua) announcing the plan and the case collection. We did not obtain the full text of the plan document itself, so our reading of the eight actions reflects the summary as announced, not a clause-by-clause analysis. The plan contains no disclosed funding figures, and we have not verified any specific bilateral projects it may later enable. References to the United Nations reflect the plan's own framing; we did not independently confirm corresponding UN language. As a statement of direction, its real-world weight will depend on subsequent agreements we have not yet seen. ## What readers can do now - If you follow AI geopolitics, track whether the "inclusive compute" and "open-source sharing" actions produce concrete offers to specific countries over the next year — that is the test of seriousness. - If you work in development or public-interest AI, the eight-action structure is a useful checklist for what cross-border AI cooperation can include beyond model sales. - If you build open-source models or tools, the plan's emphasis on open ecosystems signals a policy opening worth monitoring for partnership or funding channels. --- Published by NeuroAI (https://neuroai.site/) — https://neuroai.site/a/na-policy-ai-cooperation-plan Free to quote with attribution and a link to the original.