Every Major AI Leader Has Now Weighed In on Dario Amodei's 'Slow Down AI' Essay

Key points

• Anthropic CEO Dario Amodei published an essay on Saturday calling for slowed, better-monitored AI development, with Anthropic committing immediately to independent safety audits.
• OpenAI CEO Sam Altman and Google DeepMind co-founder Demis Hassabis publicly agreed with Amodei's direction within days of the essay appearing, as did Elon Musk.
• US President Donald Trump pushed back sharply on Monday, calling the slowdown push a "sick conspiracy" and arguing the only guardrail AI needs is a "strong president".
• Vice President JD Vance described AI companies seeking regulation as a "trojan horse", while House Speaker Mike Johnson called for a roundtable between Congress and AI companies.
• Senator Bernie Sanders backed a separate call to pause AI development altogether.

A single essay from the CEO of an AI company usually lands quietly. This one didn't.

On Saturday morning, Anthropic CEO Dario Amodei published a long piece titled "We Must Pace the Frontier", laying out why the world's most powerful AI labs should slow down. Within days, the heads of OpenAI, Google DeepMind, and SpaceX had all weighed in publicly, and so had the President of the United States. We've been following Amodei's push since July; our 13 September story "Sam Altman and Elon Musk back Amodei's call to slow down AI before a swarm takes over the internet" caught that rare show of agreement among rival lab chiefs before it became the week's defining story.

What exactly is Amodei asking for?

He wants three things. First, companies like Anthropic should invite outside auditors, independent third parties with no stake in the outcome, to check whether they're actually following their own safety rules and to report problems when they spot them. Second, the leading AI labs in democratic countries should coordinate on shared standards and limits. Third, democratic and authoritarian governments alike should try to agree on how fast AI develops globally, though Amodei acknowledged that last one is a long shot.

Critically, Amodei drew a clear line: pacing "does not mean halting model training or technical progress." It means taking enough time to check the work before shipping it.

Anthropic is committing to that first step right now, he said.

Who agreed, and who pushed back?

The agreement came fast. Sam Altman, CEO of OpenAI, the company that makes ChatGPT, said he agreed and that independent evaluators were a "great idea". On Sunday, Altman added that OpenAI welcomes a federal framework setting safety requirements across the industry, and echoed Amodei almost word-for-word: pacing doesn't mean stopping.

Demis Hassabis, who co-founded Google DeepMind and now chairs it, called the essay "the right path forward" and pointed to a standards-body framework he'd already published in July. Elon Musk reposted Amodei's essay with the words "Dario is right."

Then came the politicians, and the tone shifted.

Respondent	Position	Short take
Sam Altman, OpenAI	Agrees	Wants federal safety standards and independent auditors
Demis Hassabis, Google DeepMind	Agrees	Called the direction "correct for this critical moment"
Elon Musk, SpaceX	Agrees	Two-word endorsement: "Dario is right"
President Donald Trump	Disagrees	Called it a conspiracy; said a strong president is the only guardrail needed
VP JD Vance	Sceptical	Described industry calls for regulation as a potential trojan horse
Speaker Mike Johnson	Cautious	Wants a roundtable but ruled out a full moratorium
Senator Bernie Sanders	Agrees	Backed a separate call to pause AI development entirely

President Trump used a Monday Truth Social post to call the slowdown push a "sick conspiracy" and warned that China benefits every time American AI companies hesitate. Vice President Vance found it "a little bit weird" that powerful tech companies would ask the government to regulate them, suggesting it might be a tactic to lock out smaller competitors.

House Speaker Mike Johnson took a middle line, telling CNN he wants AI company leaders and Congress in the same room to find "the right balance", but ruled out any moratorium because of the competition with China.

Here's the part worth watching: the people running the labs actually agree on more than they disagree. The fight is between the labs and the current White House, not between the labs themselves. That gap will make any kind of federal framework very hard to build quickly.

What does this mean for ordinary people?

Nothing changes today. No new rules exist yet. What this week showed is that the companies building the most powerful AI tools publicly want someone to check their homework, even if Washington isn't ready to assign the graders.

If the outside-auditor idea gains traction, it could eventually mean more transparency about what AI systems can and cannot do before they reach your phone or your doctor's office. For now, watch whether Anthropic's self-imposed independent audit commitment actually produces public reports.

Common questions

Does "pacing" mean AI development will actually stop?

No. Both Amodei and Altman were explicit: slowing down means taking time to test and verify safety before releasing new models, not halting research.

Does this affect AI tools I already use?

Not immediately. The debate is about future standards for the most powerful, "frontier" AI systems, the advanced models that labs are still building.

Why do some politicians see this as suspicious?

Vance and others argue that big, established AI companies may support regulation partly because strict rules are easier for large companies to meet, which can squeeze out smaller rivals. It's a fair question, even if the safety concerns behind the proposals are also genuine.