"I'm not downplaying," Trump said, before making his priority explicit: "I said it from the very beginning. Whoever wins AI, and we're leading by a lot. Whoever wins AI wins." Pressed on whether he had heard directly from AI leaders, he demurred, and declined to say whether he uses the technology himself.
The Week the Doom Talk Went Mainstream
The safety debate reignited when Jacob Coxon, a 27-year-old researcher who spent three years doing pretraining work at OpenAI and Anthropic, resigned and publicly accused his industry of "gambling with our lives." The post went viral, and what followed was more striking: Anthropic's own alignment science lead, Evan Hubinger, endorsed it. "We really do earnestly believe AI could kill all humans!" Hubinger wrote on X, putting the odds at more than 10 percent within the next decade, while stressing that the risk from present models is low and that his fear is superintelligence arising from recursive self-improvement. "We do not yet have a plan to solve alignment for superintelligence and are not clearly on track to."
The candor is jarring from a company shipping increasingly autonomous products, including Claude's new ability to operate a Mac in the background, and it captures the industry's central contradiction: the people building the technology are the ones warning about it, without stopping.
China Calls It "Fearmongering"
Amodei's essay had a geopolitical edge: he argued a "Chinese lead in AI would pose grave danger for the United States and the world" and urged the Trump administration to tighten limits on AI chips reaching Beijing. China's foreign ministry dismissed the warning. "Fearmongering, confrontation and vicious competition will only disrupt the process of global AI governance which serves no one's interest," spokesperson Guo Jiakun said Monday.
That framing, safety versus the race, is exactly the tension Trump resolved in favor of the race. House Speaker Mike Johnson has likewise rejected an AI moratorium over the China threat, even as figures like Rep. Ted Lieu demand a "kill switch." For now, Washington's bet matches Silicon Valley's behavior, where personal AI agents for everyone remain the promise: build faster, worry out loud, and keep building.
Frequently Asked Questions
- What did Trump say about AI safety?
- That AI development will be "more good than bad, but by a lot," and that "whoever wins AI wins." He pushed back on calls for a slowdown days after Anthropic's CEO urged one.
- What did Anthropic's CEO propose?
- Dario Amodei called for a blanket slowdown in AI development to ensure safety, and for stricter limits on AI chips sent to China. Elon Musk and Sam Altman voiced support.
- Why did an Anthropic researcher's warning go viral?
- Jacob Coxon resigned saying the industry is "gambling with our lives." Anthropic's alignment lead Evan Hubinger agreed publicly, estimating over a 10 percent chance AI could kill all humans within a decade, while calling present-model risk low.
- How did China respond?
- Its foreign ministry called Amodei's warning "fearmongering" and said confrontation would disrupt global AI governance.
