Anthropic CEO Proposes AI Slowdown — Industry Rivals Signal Agreement

Anthropic CEO Dario Amodei is calling for the AI industry to pump the brakes. In a detailed essay, Amodei proposed a three-step plan for AI labs to collectively slow the pace of developing next-generation models, a move he argues is necessary to manage escalating safety risks. The proposal, which Amodei says “we owe it to humanity to try,” according to Engadget, has found surprising agreement from competitors, creating a rare moment of public consensus among the sector’s fiercest rivals.

The core of Amodei’s plan is to “pace the frontier,” a phrase that, as The Verge clarifies, is industry jargon for slowing down the relentless cycle of building ever-more-powerful systems. This isn’t just talk. Amodei committed to granting third-party evaluators like METR access to Anthropic’s models, a step intended to provide external validation of the company's safety commitments.

A Chorus of Caution

Perhaps the most significant part of this development is not the proposal itself, but who is agreeing with it. TechCrunch reports that OpenAI CEO Sam Altman has expressed similar views on the need to “pace the frontier.” This aligns two of the most prominent and well-funded AI labs in the world. Adding to the chorus, MarketWatch notes that Elon Musk, head of the competing lab xAI, has also backed Amodei’s call to slow down progress, framing the issue around the potential for “rogue bots.”

This public alignment is unusual. These companies are locked in a high-stakes competition for talent, computing power, and enterprise contracts. For their leaders to collectively advocate for a slowdown suggests a shared recognition that the current trajectory may be unsustainable or poses risks that outweigh the competitive benefits of speed. The consensus view across all reports is that the leaders of the AI race are, at least verbally, ready to consider tapping the brakes.

From Proposal to Practice

Amodei’s three-step plan aims to create a verifiable framework for this slowdown. While the specifics are complex, the central idea is to tie the development of future, more powerful models to the successful completion of rigorous safety evaluations. If a model fails these tests, development would pause until the issues are resolved. This moves the concept of AI safety from an internal guideline to an external, auditable process.

This is where skepticism enters the picture. While leaders at Anthropic, OpenAI, and xAI may agree in principle, the mechanics of a voluntary, industry-wide pause are fraught with challenges. The pattern indicates an attempt at self-regulation, likely to shape the narrative and pre-empt harsher, government-imposed restrictions. By proposing their own framework, these companies can position themselves as responsible actors while defining the terms of debate. The question, as TechCrunch aptly puts it, is what a slowdown would “actually look like.” In an industry driven by venture capital timelines and the promise of artificial general intelligence, a coordinated pause feels like a direct contradiction of the market’s fundamental incentives. It’s a classic prisoner's dilemma: the first company to break ranks and accelerate development could gain a significant advantage, making any voluntary pact inherently unstable.