Dario Amodei (Anthropic), Sam Altman (OpenAI), and Elon Musk jointly call for AI development slowdown to address safety concerns, breaking prior competitive silence.

For years, Dario Amodei, Sam Altman, and Elon Musk have occupied opposing corners of the artificial intelligence race. Amodei runs Anthropic, positioning the company around safety and responsible development. Altman leads OpenAI, at the centre of the commercial AI boom. Musk has repeatedly warned about AI dangers while building his own frontier AI company, xAI. Their relationships have been decidedly unfriendly: Amodei left OpenAI in December 2020 to co-found Anthropic, which became one of OpenAI's biggest rivals. Musk's relationship with Altman deteriorated into legal battle over OpenAI's direction.

Yet this past weekend, the three found something to agree on: AI development may be moving too fast.

It began with Amodei's call to "pace the frontier"—his argument that AI companies should slow the rate at which they improve their most powerful models, giving safety systems and independent oversight time to catch up. Altman quickly agreed. Musk's response was brief: "Dario is right."

This represents an unusual moment in the AI race, but the more pressing question is why these three are converging now.

**From Hypothetical Risks to Real-World Incidents**

Warnings about catastrophic AI failures have often sounded theoretical. That conversation is shifting as AI systems become increasingly autonomous. Amodei's latest warning follows incidents in which AI agents demonstrated capabilities beyond text and code generation. He specifically cited the OpenAI agents and Hugging Face incident, where a swarm of AI agents breached a third-party system.

Given the pace of progress, Amodei warned, similar systems could potentially take over large parts of the internet within six to twelve months, causing hundreds of billions of dollars in damage. Rather than calling for AI development to halt entirely, he argues companies should slow capability advancement enough to properly test and secure increasingly powerful systems.

Amodei has called for independent evaluators to have employee-level access to AI companies, industry-wide safety standards, and international coordination around frontier AI. Anthropic says it will implement the independent-evaluator measure; Altman has said OpenAI will do the same. The concept is essentially a speed limit on the frontier, not a halt.

**The Alarm from Within Anthropic**

Days before Amodei published his essay, Jacob Coxon—a researcher who spent three years on pretraining at OpenAI and Anthropic—resigned from Anthropic. In a September 9 post on X, Coxon accused both companies of "racing straight to self-improving superintelligence and gambling with our lives."

Coxon argued that increasingly capable AI systems could soon hack systems, transform entire fields, and acquire real-world power and resources. More significantly, he claimed that people building AI "earnestly believe" it could kill humanity by decade's end, arguing that some executives and senior researchers express those fears privately even if public comments are more measured. He went further than Amodei's proposal, arguing that preventing a global race may eventually require more drastic measures, including a temporary ban on improving model capabilities.

Evan Hubinger, a current Anthropic researcher, has said he personally believes there is more than a 10% chance that AI could cause human extinction within the next decade. Days later, Anthropic's CEO made a public case for slowing the race down.

**Why OpenAI Is Willing to Slow Down**

Amodei is effectively asking his biggest competitors—including OpenAI—to agree to rules constraining all of them. Pressure on OpenAI is also shifting. Following Amodei's proposal, Altman backed the idea of independent evaluators and said OpenAI would adopt the same approach, signalling willingness to coordinate with rivals on AI safety. OpenAI chief scientist Jakub Pachocki separately argued for coordinating to slow future AI development when needed.

**Elon Musk's Position**

Musk has spent years warning that AI poses an existential threat to humanity, yet he is simultaneously running xAI, competing directly in the frontier AI race. A co-founder of OpenAI who later left, he subsequently sued OpenAI and Altman over what he argues is a departure from the organization's original "not-for-profit" mission.

The three men disagree on much regarding AI, but they appear to be converging on the idea that the capability race cannot continue indefinitely without stronger mechanisms to check what these systems can actually do.

**Can the Industry Actually Slow Down?**

Every major AI company competes for customers, computing power, developers, and capital. A more capable model means more users, more enterprise contracts, and more revenue. Asking rival companies to collectively slow down creates an obvious problem: what if one doesn't?

Amodei acknowledged this dilemma in his essay, arguing that the US cannot slow AI development so much that China catches up. If "unpaced" Chinese projects pull ahead, he warned, it could create significant national security risk. His proposed answer is coordination—first among US AI companies and democratic countries, eventually through broader international agreements—while preserving the US lead over China through tighter controls.

**Trump's Resistance**

US President Trump has shown little appetite for a major AI development slowdown. On Sunday, he dismissed concerns raised by AI executives, saying the US is "leading China in AI." He called critics of AI "very negative forces" raising scenarios that will not happen, adding he wanted to ensure the US remains the industry leader.

**China's Response**

China's Foreign Ministry said on Monday that all countries should work together on AI, while warning that "fearmongering, confrontation and vicious competition" would disrupt global AI governance. Chinese officials rejected US allegations that Chinese developers engaged in aggressive efforts to extract capabilities from advanced American models through "distillation." China's Commerce Ministry called the claims groundless and accused Washington of pursuing a monopoly over the AI industry.

This puts Beijing and Amodei in an awkward position: he calls for cooperation with China on frontier AI safety while simultaneously arguing that the US should restrict China's access to advanced chips and other AI capabilities. The state-run Global Times described Amodei's essay as a veiled call to contain China and, essentially, a "Cold War playbook" for the AI sector.

**The Deeper Dilemma**

The disagreement reflects a fundamental tension. Both Washington and Beijing want to prevent the other from establishing a technological chokehold while seeking access to the other side's AI advances. This makes the politics of slowing AI considerably harder than the technology debate itself. Companies may want to slow down enough to make increasingly powerful systems safer. Governments, meanwhile, do not want to slow down so much they lose the AI race.

It may explain why three of the industry's most unlikely allies are suddenly agreeing the AI race needs brakes. The fight now is over who gets to decide where those brakes should be.