Trump and House Republicans Tell AI Chiefs to Keep Their Foot on the Gas
Anthropic's CEO called for pumping the brakes on AI development. The White House and Congress said no thanks, and pointed east.
Key points
- Anthropic CEO Dario Amodei published an open letter on 12 September 2026 calling for a slowdown in AI development.
- OpenAI's Sam Altman and Elon Musk both voiced public support for Amodei's proposal.
- Demis Hassabis of Alphabet's Google DeepMind offered tentative backing.
- President Donald Trump and House Speaker Mike Johnson opposed the idea, arguing a pause could let China overtake the US in AI.
- Johnson told CNN that rushing AI regulation could be a "national security threat."
On 12 September 2026, Anthropic CEO Dario Amodei published a lengthy open letter calling on the AI industry to "pace the frontier," meaning slow the race to build ever-more-powerful AI systems. A rare moment of public caution from someone actually running that race.
What made it stranger: his rivals largely agreed. Sam Altman and Elon Musk posted support on X. Demis Hassabis of Alphabet's Google DeepMind, whose lab competes directly with Anthropic, offered tentative backing, as first reported by The Verge AI. We covered Altman and Musk's reasoning the same day in our earlier story on the internet-swarm warning.
So why isn't anything slowing down?
Washington said no. Trump and Johnson both rejected the idea quickly, and their reasoning was identical: China.
Trump put it plainly. "Look, we're leading China in AI," he said, according to the Financial Times, "and, frankly, I want to keep it that way, because whoever wins AI, wins." Johnson echoed that on CNN, telling anchor Jake Tapper that rushing restrictions onto AI development could itself be a "national security threat."
"If Congress just races in and does some sort of emergency session to try and regulate AI, we will lose the race to China," Johnson said.
The concern has logic behind it. China's government backs its own AI labs heavily, and a US pause with no equivalent move from Beijing would give those labs breathing room. The problem is that "China might catch up" has become an all-purpose stopper for almost any AI safety conversation in Washington, and it works every time.
The real story here is the inversion. The people building the most powerful AI tools are asking for a breather; the people running the government are telling them to keep sprinting. That's not how this normally goes, and it's the detail worth watching.
What does this mean for ordinary people?
Nothing changes tomorrow. No pause is coming, and no emergency law is in the works. But when the founders of Anthropic, OpenAI, and Alphabet's AI division all think development needs slowing, that's a signal the technology has moved into territory even its creators find uncomfortable. Governments decide what happens next with that signal, not individuals.
For now, the US position is full speed ahead.
Common questions
What does "pacing the frontier" actually mean?
It means voluntarily slowing how fast labs develop their most powerful AI models, giving safety research and regulation time to catch up. Not a shutdown; more like agreeing not to sprint.
Could the US government force AI companies to slow down?
In theory, Congress could pass laws restricting certain kinds of AI development, but Johnson's comments make clear that isn't the current plan. Development decisions sit largely with the companies themselves.
