The image Xi Jinping chose was music. "The development of artificial intelligence should not be a solo performance by any single country," he told the opening of the World Artificial Intelligence Conference in Shanghai on Friday, "but rather a symphony of global cooperation." It was the first time Xi had attended the flagship event since its launch in 2018, and the choreography around him made the point plainly: the day before, 29 countries had signed an agreement to create the World Artificial Intelligence Cooperation Organization, an intergovernmental body headquartered in Shanghai, with Russia, Pakistan, Kazakhstan, Indonesia and Laos among its founding members and the UN Secretary-General in the room.
The message was pitched squarely at the developing world. Xi promised 5,000 AI training opportunities over five years, cooperation with the African Union, the Arab League, ASEAN and the BRICS bloc, and free access for 30 countries to a Chinese-built weather warning system. Cheaper models, distributed openly, are already doing the diplomatic groundwork: Chinese firms supplied 20 of the 50 most-used models on the developer platform OpenRouter in May, up from five at the start of last year. To a government in Jakarta or Lagos weighing the cost of American systems, an open Chinese model that runs for a fraction of the price is a persuasive argument, and Beijing knows it.
Woven through the speech was a running complaint about Washington, never named. Xi urged the world to "jointly oppose overstretching the national security concept in the field of AI, or placing one country's security over that of others," a direct swipe at the export controls that have blocked China from the most advanced chips. He also called for AI to remain "always under human control," pairing the openness pitch with a bid to write the safety rules too. The subtext, as one analyst put it, is that China missed its chance to shape the rules of the early internet and does not intend to miss this one.
The split screen with Washington was almost too neat to be accidental. As Xi addressed the conference, President Trump was in Washington accusing Beijing of illicitly acquiring 220 million American voter files as part of a broader effort to influence US elections, a claim China denies. One leader cast his country as the responsible custodian of a technology for all humanity; the other cast the same country as a hostile intelligence operation. Both descriptions were aimed at the same watching audience of undecided governments.
The new organization is best understood as a mirror. Late last year the United States launched Pax Silica, a framework knitting together allies including Japan, Britain, Australia, India and Israel around trusted AI supply chains. WAICO is Beijing's answer: its own bloc, its own headquarters, its own standards. The problem, analysts caution, is that mirrors do not attract the other side's reflection. No major Western democracy is likely to join a body China effectively controls, and one Washington consultant argued the real American task is not to counter-program the summit but to build a credible bilateral channel with Beijing on frontier governance, a channel both sides agreed to open after Trump and Xi met in May.
There is a deeper worry beneath the applause. Western analysts note that a China setting the global norms of AI would be a China exporting the norms of its own tightly controlled information environment, embedded in the models and the rules that govern them. Xi's symphony is a genuine offer and a genuine strategy at once. The countries deciding whether to take a seat in the orchestra are, in effect, choosing whose sheet music the next decade of AI will be played from.