5 hours ago
32
A rare moment of agreement broke out among some of Silicon Valley’s biggest rivals this week, when Anthropic, OpenAI, and Google DeepMind leaders all signaled support for slowing down artificial intelligence development. But in Washington, the reaction was far less enthusiastic. President Donald Trump and House Speaker Mike Johnson pushed back hard, framing any pause as a dangerous gift to Beijing in the ongoing US China AI competition. The clash lays bare a widening gap between the people building the most powerful AI systems on Earth and the politicians who decide how, or whether, to regulate them.
Key takeaways
- Anthropic CEO Dario Amodei published an open letter calling to “pace the frontier” and slow the pace of AI capability gains.
- OpenAI’s Sam Altman, Elon Musk, and Google DeepMind’s Demis Hassabis all voiced public support for Amodei’s proposal.
- Trump said the US is “leading China in AI” and wants to keep it that way, rejecting calls to slow down.
- Mike Johnson warned that rushing AI restrictions through Congress could become a “national security threat” and hand China the advantage.
- Amodei himself acknowledged that China’s willingness to follow any slowdown is the “toughest dilemma” facing his plan.
AI Industry Leaders Advocate Slowing Down AI Development
The push for restraint came directly from the top of the industry, not from outside critics. That distinction matters, since it’s rare for company chiefs racing to build ever-more-capable systems to publicly ask for the brakes to be applied.
Anthropic CEO Dario Amodei’s Proposal
Amodei published a lengthy essay titled “We must pace the frontier,” arguing it was time to slow how quickly AI companies improve their most advanced models. The essay followed stark warnings from Anthropic’s own safety researchers about the technology’s growing potential for catastrophic harm, including one researcher who quit over fears that Anthropic and OpenAI were, in his words, gambling with people’s lives.
Amodei’s plan outlined three steps. Anthropic has already unilaterally committed to the first: giving third-party evaluators employee-level access to verify safety practices and report incidents. The second step calls on leading AI companies in democratic countries to coordinate on common safety standards, while the third envisions cooperation between democratic and authoritarian governments — a goal Amodei admitted may be difficult to reach given how much military advantage is at stake for whoever pulls ahead.
Support from OpenAI and Alphabet Executives
Sam Altman, whose relationship with Amodei has been strained since the Anthropic CEO left OpenAI to start his own company, nonetheless backed the proposal publicly, saying he agreed on the need to pace the frontier. Elon Musk, who leads both Tesla and SpaceX, also voiced support, writing that “peer review of AI by competitors is the right way to start this off.” Alphabet’s Demis Hassabis offered a more measured endorsement, saying the essay “points towards the right path forward” even if the details still need working through.
Not everyone in the industry embraced the idea without caveats. David Sacks, who previously served as Trump’s AI and crypto czar and is a longtime critic of Anthropic, argued that if the companies truly believe their models are dangerous enough to warrant slowing down, they should simply do it rather than demand new regulatory frameworks as the price of responsibility. Meanwhile, Meta’s AI chief Alexandr Wang said the company is “rapidly scaling up” the share of its efforts devoted to alignment as its models grow more powerful, without directly addressing Amodei’s essay.
US Political Figures Oppose Slowing AI Progress
While AI executives debated the merits of pacing themselves, the political response in Washington took a sharply different tone. Both Trump and Johnson framed any slowdown as a strategic risk rather than a safety win.
Donald Trump’s Position on AI Leadership
According to the Financial Times, Trump dismissed the idea of easing off AI development, saying: “Look, we’re leading China in AI… and, frankly, I want to keep it that way, because whoever wins AI, wins.” The comment reflects a broader belief within the administration that AI supremacy is inseparable from national power, and that any hesitation could cost the US its lead.
House Speaker Mike Johnson’s National Security Concerns
Johnson echoed that view on CNN, telling anchor Jake Tapper that rushing to impose restrictions on AI development could itself become a “national security threat.” He urged calm rather than panic, warning: “If Congress just races in and does some sort of emergency session to try and regulate AI, we will lose the race to China.” His comments suggest that, at least among some Republican leaders, the perceived cost of falling behind outweighs concerns raised by the very people building the technology.
The Central Concern: Maintaining US Competitiveness Against China
Beneath the disagreement over pacing sits a single, unresolved question: what happens if the US slows down and China doesn’t? Amodei addressed that tension head-on, telling CBS News’ “Sunday Morning” that the “toughest dilemma” of his proposal is the possibility that adversarial nations choose not to follow suit. “The incentives to pull ahead and the military advantage that you get from that are so large,” he said, adding that he wasn’t sure a coordinated slowdown was even possible, “but we should try.”
The stakes are heightened by recent momentum from Chinese AI startups, whose models have gained ground against leading American offerings, stoking concern among lawmakers and Silicon Valley figures alike. Days after Amodei’s essay, Chinese President Xi Jinping called for a “consensus-based global AI governance framework” at a BRICS summit in India and pledged China would help build an open-source AI community within the bloc. Trump is scheduled to meet Xi at the White House later this month, with AI expected to be a central topic of discussion.
This is the core of why the debate matters beyond Washington politics. If US policymakers treat any voluntary industry slowdown as a competitive liability, safety-driven proposals like Amodei’s may struggle to gain traction regardless of how much support they get from the companies actually building the technology. That leaves the industry in an unusual position: the people closest to the risks are asking for restraint, while the people who set the rules are betting that restraint itself is the bigger danger.
FAQ
Why is Anthropic CEO Dario Amodei calling to slow down AI development?
Amodei proposed slowing AI progress to “pace the frontier” and ensure safety keeps up with advancing capabilities, warning that unchecked development carries growing risks.
What is Donald Trump’s stance on slowing AI development?
Trump opposes slowing AI development, emphasizing that the US must maintain its lead over China, saying “whoever wins AI, wins.”
What national security concerns does House Speaker Mike Johnson raise about AI regulations?
Johnson warns that rushing AI restrictions through Congress could become a national security threat and cause the US to lose the AI race to China.
Do AI industry leaders and US political figures agree on how to handle AI development speed?
No. Industry leaders like Amodei, Altman, Musk, and Hassabis have voiced support for slowing AI progress for safety reasons, while political leaders like Trump and Johnson prioritize maintaining US leadership in the broader US-China AI competition.
Article produced with the assistance of artificial intelligence and reviewed by the editorial team.
English (US) ·