View / White House is isolated in brushing off AI safety
The White House and its allies are lonely in their insistence that AI development in America must accelerate without governments setting ground rules.
At the UN General Assembly in New York City, as well as in meetings with Chinese leader Xi Jinping in Washington this week, President Donald Trump’s team was clear: US companies are the leaders in AI; they plan to keep it that way; and they reject any international attempts to rein in the technology.
Trump, with his call to leave AI “exactly where it is” contrasts with Xi, who said responsible AI development is necessary — as other Chinese officials said this week that they support a global governance framework.
Without official support from the US, however, any such agreement will be difficult to execute. White House science and tech adviser Michael Kratsios rejected global efforts in a UN address that he called an “intervention.”
As a result, nations may be left setting nonbinding rules and best practices without the buy-in of the biggest companies’ home country. Meanwhile, frontier labs are pushing forward with their own self-regulatory schemes, which will ask them to agree on testing standards in a highly competitive environment.
Yet even AI CEOs are more aligned with the UN than with Trump. Companies are pushing for coordinated “AI pacing,” saying they support global efforts to coordinate on safety and affirming that they cannot set the standards for responsible AI use on their own.
“The international community will need to establish a common framework to manage the risks associated with frontier AI models,” said Jean-Noël Barrot, the French official who chaired the UN National Security meeting on AI featuring the CEOs of OpenAI, Anthropic and Hugging Face.
Polling shows Americans are largely skeptical of claims that the US must beat China at AI development, and that people are generally worried the technology is moving fast enough to pose a risk to their jobs and livelihood. Those surveys show worries about AI acceleration across party lines.
Trump’s AI isolation from the world, and much of the electorate, puts congressional Republicans in a strange position, too. They’re advocating for AI safety and slowdowns and sponsoring bipartisan legislation to regulate AI, while their party leader presses the gas.
As Semafor first reported on Thursday, Sen. Todd Young, R-Ind., has asked Secretary of State Marco Rubio to convene National Security Council meetings with cyber experts and AI developers about potential national security threats, including possible AI-enabled attacks on critical infrastructure.
“The United States should continue to lead the world in the development and deployment of artificial intelligence,” Young wrote to Rubio. “Maintaining that leadership also requires ensuring that our government keeps pace with emerging capabilities and understands the risks they may present.”
Trump’s allies often see ulterior motives behind much of the current advocacy for stricter AI safety rules. Some argue that companies as big as OpenAI, Google DeepMind, and Anthropic are seeking “regulatory capture” by trying to shape legislation that only their companies will have the means to comply with, therefore giving themselves a leg up.
Others decry “doomerism,” or the belief that AI could kill off humanity or eliminate jobs, as hysterical and speculative. Criticizing the effective altruism community, which many AI safety groups have connections to, is another burgeoning White House tactic.
But the core of the president’s message is simple: The US should do nothing that might give China, with its cheaper open-source models, an advantage in the AI race.
