U.S. and China to Hold First AI-Exclusive Safety Talks Under Trump Administration

The United States and China are preparing for a high-level AI safety dialogue scheduled for mid-September 2026. The talks represent the first official bilateral discussion devoted exclusively to artificial intelligence since the beginning of President Trump's second term.

Led by U.S. Treasury Secretary Scott Bessent, the delegation aims to address escalating global risks associated with frontier AI. Primary concerns for the dialogue include the threat of autonomous AI-driven cyberattacks and the practice of "distillation," in which Chinese firms are accused of using outputs from U.S. models to train their own systems. These discussions come as both nations navigate a precarious balance between technical competition and the prevention of systemic failure.

The Push for 'Pacing'

While governments negotiate, a significant movement has emerged from within the industry. In July 2026, more than 1,200 employees from the world's leading AI laboratories—including OpenAI, Anthropic, Google DeepMind, and Meta—signed an open letter titled "Pacing the Frontier." The signatories are urging the U.S. government to coordinate international efforts to moderate the speed of frontier AI development, specifically targeting the acceleration of automated AI research.

The letter reflects a growing anxiety among researchers that capabilities are advancing faster than the safety frameworks required to contain them. By calling for a coordinated "pacing" of the frontier, these employees are seeking a diplomatic and regulatory mechanism to ensure that safety measures keep pace with the raw power of the models being deployed.

A Fragile Regulatory Balance

This diplomatic push occurs alongside the adoption of the "Carolina Principles," a framework signed by the U.S., China, and G-20 ministers. These principles are designed to discourage the implementation of overly restrictive, AI-specific regulations that could stifle innovation. However, the tension remains high; China has publicly warned of "extreme AI loss of control risks," even as it seeks to avoid the very regulations that might mitigate those risks.

Why Coordination Matters

As frontier AI capabilities reach a critical tipping point, the ability of the two primary AI superpowers to coordinate on safety is no longer just a matter of policy, but of global stability. The emergence of autonomous agents capable of hijacking websites and forging logs suggests that the risks of AI-driven warfare transcend traditional geopolitical rivalry.

"We are at a tipping point where frontier agents can cause massive damage when unmonitored," said Samm Sacks, a senior fellow at New America. "Both the U.S. and China are vulnerable. This is beyond geopolitical rivalry."

What to Watch

Observers will be watching the mid-September talks to see if the U.S. and China can move beyond general principles toward concrete agreements on monitoring autonomous agents. Whether the U.S. government will adopt the "pacing" recommendations from the 1,200 industry employees remains unconfirmed, but the alignment of internal industry warnings and high-level diplomatic urgency suggests a shift in how the AI frontier is managed.