The first two powers of artificial intelligence do not see its future in the same way. During the summit of the Asia-Pacific Economic Cooperation (APEC), organized at Shenzhenin China, Donald Trump And Xi Jinping exposed the November 19, 2026 two opposing approaches to AI governance. The American president rejected any international regulation, the Chinese president pleaded for cooperation on security. Between the two, discussions on a common incident management mechanism continue, without a published agreement.
Trump: “We oppose any globalist regulation”
Before APEC leaders, Donald Trump reaffirmed his line. “It’s a superintelligence, but we have to monitor it,” he said, according to Euronews, using his administration’s preferred term for AI. “We have to be careful. It’s almost all good, but there are bad elements, and we have to be very careful in dealing with those bad elements. »
However, he was categorical on the framework: “We oppose any form of globalist regulation or anything that would seek to control and stifle superintelligence, and we will not let that happen. » The American president has regularly rejected calls for international governance of AI, believing that it would slow down the United States in its race against China.
Xi: keeping AI “under human control”
According to the official Chinese news agency new china (Xinhua), Xi Jinping responded that the two countries should cooperate. “China and the United States need to strengthen communication and coordination to guard against the security risks of artificial intelligence and ensure that artificial intelligence always remains under human control,” he said, according to the agency.
This speech is not new. Xi Jinping had already affirmed this summer that AI must remain “always under human control”, calling for more laws and regulations, technological surveillance, early warning systems and emergency response. During APEC, he also invited leaders to “deepen cooperation in the digital economy and artificial intelligence”, without further details.
| Position | United States (Donald Trump) | China (Xi Jinping) |
|---|---|---|
| International regulation | Rejection of any “globalist regulation” | Call for coordination between the two countries |
| Priority displayed | Do not slow down innovation in the face of China | Keeping AI under human control |
| Recognized risks | “Bad elements” to be managed with caution | Security risks to prevent together |
An emergency channel for incidents
Despite their differences, Washington and Beijing are moving forward on a concrete point. In September 2026the two countries agreed to create a emergency communication channel dedicated to AI incidentsthe terms of which were to be discussed in Shenzhen. According to US News, US Treasury Secretary Scott Bessent indicated that discussions would focus on communication protocols in the event of a security incident and the exchange of best practices for testing advanced models.
However, at the end of the discussions, no common framework or joint rule had been published, according to the available reports. Messages on social media claiming that the two leaders had agreed on “reasonable rules” are not confirmed.
Why an Incident Channel Matters
Such a channel may seem modest given the challenges, but it meets a very concrete need. Imagine that an AI agent developed in one of the two countries disrupts the other’s infrastructure, or that an AI-assisted cyberattack is wrongly attributed to one state. Without a rapid and recognized means of communication, such an incident could be interpreted as a hostile act and cause an escalation. A dedicated channel would verify facts, report an incident in progress, and coordinate a response.
The exchange of best practices on advanced model testing is also of interest: the two countries are developing the most powerful systems in the world, and a shared understanding of risk assessment methods, particularly in cybersecurity and biology, could limit slippage. It remains to be seen to what extent technological rivalry will give way to this minimal cooperation.
Shared risks despite rivalry
Euronews recalls that, despite technological rivalry, several risks worry the two powers: infrastructure hackingthe use of AI to help make biological weaponsand the uncontrolled AI agentsthese tools that accomplish tasks online with little human intervention. OpenAI agents have notably broken into sites, including an Australian government portal and several American government sites, incidents followed by similar reports from other laboratories. THE military uses and the mass surveillance Other points of friction remain.
The meeting took place against a backdrop of tensions around chips. Three days earlier, Washington imposed a license on five countries to block China’s remote access to AI chips. The recent resignation of Jensen Huang from the management of Nvidia also weighed on trade.
And Europe in all this?
The European Union defends a third way: that of a restrictive regulation based on risks, with the European regulation on AI, while seeking not to lose out in the technological race. This position is increasingly contested internally: according to the press, a senior German government official pleaded in mid-November for European rules more favorable to innovation.
The divide between Washington and Beijing complicates European ambitions for global governance. Multilateral initiatives are struggling to bring the two powers together: at COP31, neither the United States nor China signed the Antalya commitment on AI and the climate. Canadian Prime Minister Mark Carney summed up the concern of many countries at APEC: according to him, the lack of global governance of AI represents a “huge risk”.
What will play out
The Shenzhen summit confirms that AI governance will first be negotiated between Washington and Beijing, on a case-by-case basis, rather than in a major international treaty. The emergency channel on incidents, if it sees the light of day, would be the first concrete mechanism for dialogue between the two powers on this subject, comparable to the hotlines established during the Cold War to avoid nuclear escalations. For Europe, the challenge is to remain audible: by defending its rules, developing its own capabilities and establishing alliances, such as the technological partnerships concluded with the United States as part of the Pax Silica initiative and the San Francisco Declaration.

