AI Leaders Call for a Paced Approach to Frontier Development

The CEOs of Anthropic and OpenAI, Dario Amodei and Sam Altman, have issued a joint call for the industry to moderate the current pace of frontier AI development. The leaders argue that it is essential for safety protocols and oversight measures to evolve in lockstep with the rapid advancements of increasingly capable models.

“We must slow the pace at which we improve the capabilities of AI models,” Amodei stressed in a recent essay, pointing out that AI progress has accelerated drastically due to the effects of recursive self-improvement. “Left unchecked, it could outrun our ability to understand and control these systems,” he warned. The Anthropic chief referenced a July incident involving an OpenAI-related breach at Hugging Face, suggesting that even more capable future systems could potentially cause catastrophic damage if not properly governed.

“Given the accelerating rate of AI capability development, it’s my worry that in six-12 months such a swarm could be capable of taking over the entire internet with a persistent botnet,” Amodei cautioned. To address these risks, he proposed a three-step framework that includes the use of permanent, third-party evaluators with deep internal access to AI firms, the establishment of common safety standards among developers in democratic nations, and broader global regulatory coordination.

OpenAI CEO Sam Altman voiced his support for the proposal via a post on X, confirming that his company would also implement the use of independent evaluators. “The world deserves confidence that American companies developing increasingly capable AI will act responsibly,” he said. Altman further clarified the intent behind the move, stating, “When we talk about ‘pacing’, we do not mean ‘stopping’. No amount of American competitive pressure should justify recklessness, or let capabilities get ahead of alignment and monitoring.” The proposal also received a brief endorsement from SpaceX CEO Elon Musk, who posted: “Dario is right”.

These calls for restraint come amid growing global anxiety regarding the behavior of autonomous systems. Just last week, Anthropic reported that threat actors had attempted to misuse its Claude models for surveillance, cyberattacks, and weapons development. Additionally, Anthropic researcher Evan Hubinger recently issued a dire warning, suggesting there is a greater than 10% chance that advanced AI could lead to the extinction of humanity within the next decade.

However, the proposal for a slowdown has met with significant political pushback in the United States. President Donald Trump rejected the calls, telling reporters: “I think you have a lot of very negative ⁠forces that are bringing it up that shouldn’t be bringing it up, and they’re bringing up things that won’t happen. We’re leading China in AI. We’re the most sophisticated country in the world, and frankly I want to keep it that way because whoever wins AI wins.”

The Chinese government has also voiced its opposition to the framework. Foreign Ministry spokesperson Guo Jiakun advocated for an open and benevolent approach to AI, warning that fostering perceived threats and engaging in malicious competition would disrupt global governance. Furthermore, the state-backed Global Times published an editorial characterizing Amodei’s proposal as a “Cold War playbook” designed to hinder China’s technological development and preserve Western geopolitical dominance.