Xi Jinping stood in front of more than a thousand Chinese tech firms on Friday and declared that artificial intelligence 'should not be a solo performance by a single country.' This was the same week China quietly established a new international AI governance body headquartered in Shanghai, with 29 member countries and zero Western nations. The audacity is genuinely impressive.
The Speech Nobody Could Say With a Straight Face
At the opening of the World Artificial Intelligence Conference in Shanghai, Xi called for AI to be developed as 'a symphony of international cooperation.' He said countries should 'jointly oppose overstretching the national security concept in the field of AI.' He said no single country's security should be placed over that of others.
This is China's president. Speaking at a Chinese conference. One day after China launched its own intergovernmental AI cooperation group. Without inviting the United States or European Union. The speech was reported by AFP via Channel News Asia, and you have to respect the sheer confidence it takes to deliver it.
Xi also called for AI to 'always be under human control' and pushed for a 'people-centric' approach to the technology. Both of those things sound reasonable in isolation. In context, it's a country with extensive domestic surveillance infrastructure telling the world it wants a seat at the table on global AI ethics. Okay. Sure. Pull up a chair.
The New Governance Group Nobody in the West Joined
Here's where the real news is buried. On Thursday, a day before Xi's big speech, Chinese foreign minister Wang Yi and representatives from 29 countries agreed to establish the World Artificial Intelligence Cooperation Organization, which will be headquartered in Shanghai. Members include Russia, Pakistan, and Indonesia. The stated goal, according to state media, is to ensure the 'healthy and orderly' development of AI.
The United States was not there. The European Union was not there. When an AI consultant quoted by AFP pointed this out, he offered a generous interpretation: Europe already has its AI Act and the US is defining its own rules, so of course they stayed away. That's one way to frame 'we weren't invited and wouldn't have come anyway.'
New York University professor Arun Sundararajan told AFP that while there were 'small glimmers of recent cooperation between Presidents Xi and Trump,' it was 'hard to imagine there being a single approach to AI governance globally.' That is a very polite academic way of saying the world just split into two AI governance blocs and nobody is pretending otherwise.
China's AI Is Not Playing Around Anymore
The WAIC conference is also a showcase, and what China is showing off matters. According to CNA's reporting, around 3,000 products are on display at this year's event, ranging from high-end semiconductor systems to a smartphone capable of autonomously operating apps.
Early Friday, Beijing-based startup Moonshot AI released a new flagship model called Kimi K3, which it described as delivering 'frontier-level performance.' Huawei showed its Atlas 950 supernode, a major AI computing architecture. MiniMax unveiled its M3 model. This is not a country playing catch-up in a lab somewhere. This is a public demonstration of scale.
And the adoption numbers are staggering. Daily consumption of AI 'tokens' in China has increased a thousandfold over the past two years, according to state media citing officials at the conference. Companies like Siemens are reportedly choosing Chinese open-source AI models over American alternatives because they're cheaper, more customizable, and perform comparably. That last detail should make American AI executives uncomfortable, and it probably does.
America's AI Companies Are Not Exactly Covering Themselves in Glory Either
To be fair, the US has handed China some genuine rhetorical ammunition. According to CNA's reporting, both OpenAI and Anthropic had to temporarily delay releasing their latest AI models after government concerns that the systems could help hackers break into critical infrastructure. Read that again. The government told the companies to hold their models. The companies complied. The debate about who controls the most powerful AI technology is not just happening in Beijing.
An AI researcher named Mike Luan, speaking to AFP outside the WAIC venue, put it plainly: 'Models these days are controlled by very few.' He argued that more cooperative governance could expand access. That is not a radical statement. It is a factual one. The most powerful AI systems on earth are controlled by a handful of American companies, and the US government is increasingly dictating their release schedules. Xi is not wrong that this is a concentration of power. He's just not the guy you'd pick to fix it.
What This Conference Actually Signals
Shengyun Lu, an AI entrepreneur and founder of a Shanghai consultancy called Praxis Advisory, told AFP that China is trying to lead 'not only in terms of the technology development, but also in terms of AI governance.' He compared the regulatory challenge to nuclear power: it should be managed with the same seriousness.
That framing is revealing. Nuclear governance took decades, multiple near-catastrophes, and an entire Cold War to sort out, and it never fully sorted out. The suggestion that AI governance is on that same level of civilizational stakes is not hyperbole, and the fact that the two largest AI powers on earth are building parallel governance structures rather than a shared one is genuinely alarming.
UN Secretary-General Antonio Guterres attended WAIC. So did leaders from Cambodia and Thailand. The conference has real geopolitical weight. This is not a tech expo. It is a diplomatic move dressed up as a product launch.
The Dingo Take
Xi Jinping giving a speech about how no single country should dominate AI is a bit like a monopolist giving a keynote on the virtues of free markets. The argument is not entirely wrong. The messenger just happens to be running the world's most sophisticated state-controlled internet, a domestic surveillance system that makes the NSA look like it's working with Post-it notes, and now a new international AI governance organization that somehow forgot to invite any country with a free press.
But here is the uncomfortable part. The US response to all of this has been to restrict chip exports, delay AI model releases over hacking concerns, and let a handful of Silicon Valley companies make civilization-altering decisions with minimal democratic oversight. Washington is not exactly winning the 'trustworthy steward of global AI' argument by default. We have two flawed, self-interested powers racing to define the rules of the most transformative technology in human history, and neither one is doing it with the rest of the world's interests as the primary concern.
The world is splitting into two AI blocs. The rules are being written right now. And the main event is two governments bickering over whose concentration of AI power is more justified. Sleep tight.
