2 superpowers, 1 AI dilemma: Trump-Xi talks confront rogue AI risk

For the AI community, the U.S.-China meeting isn't just geopolitics. It could shape the technical guardrails—model evaluation standards, incident reporting norms, and red-teaming protocols—that developers in both countries will face. As China narrows the capability gap, expect safety policy to become a competitive lever as much as a cooperative one.

U.S. President Donald Trump and Chinese leader Xi Jinping are expected to meet during the week of September 20, 2026, with artificial intelligence near the top of the agenda. The Associated Press reports that the two leaders will sit down against a backdrop of rising concerns that rogue AI could pose threats to humanity, even as their governments compete to dominate the technology. The moment is defined by a paradox: the same capabilities that promise economic and strategic advantage are also the ones both nations fear could escape their control or be weaponized by the other side.

“

President Donald Trump and Chinese leader Xi Jinping are expected to meet during the week of September 20, 2026, with artificial intelligence near the top of the agenda.

The meeting comes as China continues to narrow the technology gap with the United States. That narrowing has shifted the balance of the AI race from one of clear U.S. preeminence to a more contested landscape in which both governments view the other's advanced AI systems as potential threats. The AP reports that each country has expressed concern over how AI in the other's hands could hurt them, a mutual suspicion that complicates any effort to cooperate on safety. Scott Singer, co-director of the China AI Initiative at the Carnegie Endowment for International Peace, captured the dynamic bluntly: "The U.S.-China conversation in AI is getting caught in the ups and downs in the broader U.S.-China relationship." That framing matters because it suggests that even well-intentioned safety discussions are unlikely to escape the gravitational pull of trade disputes, export controls, military tensions, and broader strategic rivalry.

The diplomatic groundwork for the expected Trump-Xi meeting was laid on Sunday, September 20, 2026, when U.S. Treasury Secretary Scott Bessent met Chinese Vice Premier He Lifeng in New York. According to the AP, the two officials discussed the setting up of a mechanism ahead of the leaders' meeting. That mechanism, if it materializes, could become one of the few institutional channels for ongoing U.S.-China AI discussions. Still, the source reporting stops short of describing the mechanism's scope, membership, or whether it would have any binding authority. In the current environment, even a modest working group focused on AI safety could be significant, but its success would depend on sustained political will on both sides—something that has been in short supply.

The stakes for the global AI ecosystem are substantial. A genuine U.S.-China safety dialogue could lead to shared norms around model evaluation, incident reporting, and dangerous capability thresholds. Such alignment would reduce uncertainty for developers, investors, and users in both countries. But if the meeting produces only vague statements, the two largest AI ecosystems may drift further toward incompatible standards, creating compliance burdens and fragmenting the global market. The security dilemma is especially acute because AI capabilities are dual-use: advances in reasoning, autonomous planning, and code generation can rapidly cross the line from commercial tools to national security concerns.