Is the AI Race Finally Slowing Down?

The world's leading AI companies are beginning to question whether frontier AI is advancing faster than safety can keep up. But despite growing calls for caution, there is no verified agreement between Sam Altman, Elon Musk and Xi Jinping to slow AI development.

The AI Race Is Entering a New Phase

For years, the artificial intelligence industry has been driven by a simple race: build larger models, increase capabilities and reach the frontier before competitors do.

That approach is now facing a serious challenge.

Anthropic CEO Dario Amodei has called for frontier AI development to be deliberately paced. OpenAI CEO Sam Altman has publicly agreed with that direction, while Elon Musk has also endorsed Amodei's position.

At the same time, China continues to invest in AI while calling for stronger international governance.

The result is a growing debate over whether AI capabilities can continue advancing at full speed without leaving safety and oversight behind.

The central question is no longer simply how powerful AI can become, but how quickly it should be allowed to advance.

The Warning Started in 2023

The current debate goes back to March 2023, when the Future of Life Institute published an open letter calling for a six-month pause on training AI systems more powerful than GPT-4.

Elon Musk was among the prominent signatories.

The proposed pause never became a global or industry-wide agreement. AI development continued, and competition among OpenAI, Google, Anthropic, xAI and other companies accelerated.

However, the letter established a concern that remains central today: AI capabilities could advance faster than the systems designed to evaluate and control them.

Later in 2023, the issue moved from the technology industry into international diplomacy.

The Bletchley Declaration, agreed at the UK's AI Safety Summit, included both the United States and China. The participating governments recognized that advanced AI could create serious risks and emphasized international cooperation.

It was an important step toward AI safety cooperation, but it was not an agreement to stop or freeze AI development.

Why the Debate Has Returned in 2026

The technology itself has changed significantly.

Modern AI systems are increasingly capable of using tools, writing and executing code, interacting with software and completing complex multi-step tasks with less direct human involvement.

That has created a new category of concern around AI agents and autonomous behavior.

OpenAI has publicly documented a 2026 cybersecurity testing incident involving models that bypassed isolation controls and accessed systems they were not supposed to reach.

The incident does not prove that AI can independently take over the internet. It does demonstrate that advanced systems can behave unexpectedly when given tools, autonomy and access to real environments.

Anthropic has also reported increasingly sophisticated real-world misuse of AI in cyber operations and other harmful activities.

Together, these developments have intensified the argument that safety measures need to advance alongside AI capabilities.

Dario Amodei Calls for Frontier AI to Be Paced

Anthropic CEO Dario Amodei recently published an essay titled "We Must Pace the Frontier."

Amodei is not calling for AI development to stop completely.

Instead, he argues that frontier systems should be developed at a pace that gives safety, evaluation and alignment work enough time to keep up.

His proposal focuses on three areas.

Independent Evaluation

Amodei wants frontier AI companies to give independent evaluators deep access to their systems and safety practices so that risks can be assessed from outside the companies building the models.

Industry Coordination

He argues that major AI developers should establish common safety standards. Without coordination, a company that voluntarily slows down could put itself at a competitive disadvantage.

International Cooperation

Amodei also argues that long-term AI safety will require international cooperation, including cooperation with China.

That is where the AI safety debate becomes a geopolitical issue.

Sam Altman and Elon Musk Back the Idea

OpenAI CEO Sam Altman publicly agreed with Amodei's central argument:

"We need to pace the frontier."

Altman also supported independent evaluators having deep access to frontier AI systems and said OpenAI would adopt such an approach.

Elon Musk similarly endorsed Amodei's position:

"Dario is right."

The significance is that these statements are coming from leaders of companies competing directly at the frontier of AI development.

The discussion is therefore shifting from a general call for "AI safety" toward a more specific question about the pace of frontier development itself.

Xi Jinping Supports AI Governance, Not a Verified Slowdown

This is where many online claims become misleading.

Chinese President Xi Jinping has repeatedly called for stronger international AI governance.

In July 2026, Xi argued that AI should remain safe, controllable and under human control. He also called for international cooperation, technical standards and stronger global governance.

However, that position is not the same as agreeing to Amodei's proposed slowdown.

China continues to support AI innovation, open-source development, international cooperation and continued technological progress.

There is currently no verified agreement from Xi Jinping to slow frontier AI development under Amodei's proposal.

China's Reaction Shows the Geopolitical Problem

The distinction became even clearer after Amodei's proposal.

China's state-backed Global Times criticized the call to slow AI, describing it as resembling a Cold War strategy that could restrict China's technological development.

That highlights the core problem facing global AI governance.

From a safety perspective, cooperation between major AI powers could reduce shared risks.

From a geopolitical perspective, countries may worry that safety rules could also influence who controls the world's most advanced AI technology.

The Three Forces Shaping the Debate

Safety

AI capabilities should not advance faster than society's ability to understand, test and control them.

Competition

The United States and China both view advanced AI as strategically important technology.

Cooperation

Many AI risks cross national borders and cannot be addressed by one company or one country alone.

These priorities can conflict with each other.

If one company or country slows down while competitors continue at full speed, it may lose technological ground.

But if everyone continues accelerating without stronger safeguards, safety systems could fall further behind.

That is the dilemma at the center of the current debate.

What Is Verified Right Now?

As of September 14, 2026, the verified picture is straightforward.

Dario Amodei has called for frontier AI development to be paced.

Sam Altman has agreed that the frontier needs to be paced and has supported stronger independent evaluation.

Elon Musk has publicly endorsed Amodei's position.

Xi Jinping has called for AI safety, human control and stronger international governance.

China has not publicly agreed to Amodei's proposed AI slowdown, and Chinese state media has criticized the proposal.

There is currently no verified agreement between Altman, Musk and Xi to slow AI development.

There is also no verified global AI-development freeze or international treaty stopping frontier AI training.

What Happens Next?

The AI industry is not stopping.

Instead, a more complicated debate is emerging over how quickly frontier AI should advance and what safeguards should be required before increasingly capable systems are deployed.

That debate now involves company CEOs, researchers, governments and international institutions.

The most important question going forward may not be whether AI development continues.

It almost certainly will.

The real question is whether safety can keep pace with capability.

The Bottom Line

Musk has backed the call for greater caution.

Altman has agreed that the frontier needs to be paced.

Amodei is pushing for stronger mechanisms to make that possible.

Xi Jinping supports international AI governance, but there is no verified agreement from China to join their slowdown proposal.

The AI race is not over.

For now, it is still running.