This is the official English translation, published by China's Ministry of Foreign Affairs, of the keynote address delivered by President Xi Jinping at the opening ceremony of the 2026 World AI Conference and High-Level Meeting on Global AI Governance in Shanghai on July 17, 2026 — the first such address by a Chinese head of state at the event. It is a primary text: China's authoritative statement of its position on global AI governance, and the document behind the widely quoted "encourage open source, openness, collaboration and sharing" line.
Framing
The speech opens by dating the field to the 1956 Dartmouth Workshop, "seventy years ago," and places AI in a sequence with the steam engine, electrification, and the Internet as technological revolutions that "profoundly reshaped our way of work and life." It poses four questions it presents as demanding answers from the international community: how to get along with thinking machines; how to ensure security when algorithms participate in decision-making; how to address ethical challenges through adaptive governance; and how to realize "AI for all when the divide keeps widening."
The stated Chinese position is that all countries should "take a people-centered approach and develop AI for the positive and for good," making AI "an important driver for shared prosperity and common security."
The four principles
First — openness and win-win. The text describes AI as "a new engine of world economic growth and an accelerator for the shift of growth drivers" that "is moving from the digital world into the physical world," and calls on states to "seize this rare, historic opportunity to encourage open source, openness, collaboration and sharing." It pairs this with coordinated advancement across three industrial tracks: transforming and upgrading traditional industries, growing emerging industries, and forward-looking planning for future industries.
Second — risk awareness, security, and controllability. AI "should be a trusted tool for humanity," and states should take seriously "the various types of inherent and secondary risks" it may trigger. The prescribed instruments are laws and regulations, technological monitoring, and early-warning and emergency-response systems, to "prevent abuses and malicious use, and ensure that AI is always under human control." The principle closes with a direct diplomatic argument: states "should jointly oppose overstretching the national security concept in the field of AI and placing one country's security over that of others" — a formulation that in Chinese diplomatic usage refers to export controls and related restrictions. See Export Controls (AI).
Third — inclusiveness and mutual learning between civilizations. AI development "should not erode or undermine the diversity of world civilizations or the uniqueness of cultures of different countries." The text calls for shaping "the values of AI with humanity's common values" and using AI to increase understanding and exchange among civilizations.
Fourth — solidarity and improved global governance. The speech calls for "true multilateralism," recognizes "the important role of the United Nations," and seeks alignment and coordination on "AI development strategies, governance rules and technical standards, so as to form a consensus-based global governance framework at an early date." It ties this to Global South capacity building, to bridging the AI and digital divides, and to preventing "creating new historical injustice in AI."
Domestic account
The address situates itself at the start of China's 15th Five-Year Plan. It describes China as having "promoted interplay between an efficient market and a well-functioning government," advanced the AI Plus Initiative, and built an ecosystem in which the core smart-economy industries are "worth at least RMB 1 trillion yuan," with intelligent manufacturing described as "another shining hallmark of Chinese modernization."
On safety it states that China is "continuously improving laws, regulations, policies, mechanisms, application norms, as well as ethical principles to make sure that AI is safe, secure and controllable," using the image of "this fine steed of AI gallop[ing] with both speed and stability." See Chinese AI Policy.
Institutional record and new commitments
The speech enumerates what China presents as its contributions of AI-related international public goods since Xi proposed the Global AI Governance Initiative: promoting consensus adoption of the UN General Assembly Resolution on Enhancing International Cooperation on Capacity Building of Artificial Intelligence; publishing the AI Capacity Building Action Plan for Good and for All; announcing the AI Plus International Cooperation Initiative; and advocating establishment of the World Artificial Intelligence Cooperation Organization.
It then confirms that WAICO "has come into being in Shanghai" — describing this as answering "the call of the Global South" — and announces three commitments for the following five years:
| Commitment | Detail |
|---|---|
| AI training | 5,000 opportunities in AI training and seminar programs for developing countries |
| Regional cooperation centers | International AI application cooperation centers with ASEAN, the League of Arab States, the African Union, CELAC, the Shanghai Cooperation Organization, and BRICS |
| MAZU | Access for 30 countries to the AI-powered meteorological warning system MAZU |
The closing passage calls for making AI oversight and governance "precise and effective," and for constantly refining "measures to forestall loss of control."
Reading
Two features of the text are load-bearing for how it has been cited. The open-source language in the first principle is a state-level endorsement of open-weight release, and is the passage Ben Thompson reads in Who's Afraid of Chinese Models? (Ben Thompson, Stratechery, July 2026) as a commoditize-your-complements strategy, emphasizing its explicit link to "the physical world" where China's robotics position benefits from widely available models. See Open-Weight Frontier Models.
The second is the pairing within a single principle of an endorsement of human control and loss-of-control prevention with an objection to "overstretching the national security concept." The speech therefore advances a safety argument and an anti-export-control argument together, rather than treating them as separate tracks.
Provenance
Retrieved from the Ministry of Foreign Affairs English service. The gap-identifier verified the text on 2026-07-19 against the Chinese-language canonical version on mfa.gov.cn (publication date 2026-07-17), and confirmed that distinctive passages — the 5,000 training opportunities, WAICO's Shanghai headquarters, the six named partner organizations, and the open-source language — match Reuters and Wall Street Journal coverage. As an official translation of a head-of-state address, it is authoritative for what was said and is not evidence for the factual claims it contains, including the RMB 1 trillion smart-economy figure.
Relationships
- supports: World Artificial Intelligence Cooperation Organization (WAICO) — confirms establishment and Shanghai headquarters
- supports: Chinese AI Policy — authoritative statement of the governing position
- supports: Who's Afraid of Chinese Models? (Ben Thompson, Stratechery, July 2026) — the primary text behind Thompson's commoditize-your-complements reading
- contradicts (partial): Export Controls (AI) — advances the "overstretching the national security concept" objection to restriction-based policy
- related: Pax Silica (US-Led AI Supply-Chain Initiative) — the US initiative positioned against the governance architecture this speech advances
- related: US-China AI Competition: Different Races, Different Metrics, Open-Weight Frontier Models, Xi Jinping