As leaders from the United States and China prepare to meet in Shenzhen next month for their third summit of the year, concerns are mounting over the absence of a clear American strategy to address the risks posed by artificial intelligence (AI). While both nations vie for supremacy in this rapidly advancing technology, experts warn that discussions on AI safety may be coming too late to prevent potentially catastrophic consequences.
The U.S. currently leads in AI development, propelled by frontier companies such as Anthropic and OpenAI. However, its approach to regulating the technology lags behind China’s more centralized and coordinated efforts. Under President Xi Jinping, China pursues a coherent national AI strategy, managed at the highest levels of government. Ding Xuexiang, China’s executive vice premier and member of the Politburo Standing Committee, recently engaged with American advisors at Tsinghua University, including former Apple CEO Tim Cook, to discuss AI policy and security issues. Additionally, Chen Yixin, head of China’s intelligence service, published an extensive analysis outlining both the threats and opportunities AI presents to national security.
Chinese officials frame AI safety within the context of protecting state interests rather than broader existential concerns often voiced in the United States. Analysts note that China aims to balance rapid AI development with measures that maintain safety while reserving the authority to intervene if political or security risks arise.
By contrast, U.S. policymakers and industry leaders exhibit a more fragmented stance. Some American AI executives, including Dario Amodei of Anthropic and OpenAI CEO Sam Altman, advocate for deliberate pacing and government oversight to mitigate risks. Meanwhile, President Donald Trump has emphasized a competitive framework, portraying AI development as a zero-sum contest with China. Treasury Secretary Scott Bessent echoes this competitive urgency but has also led efforts to open dialogue with Chinese counterparts on AI-related matters.
Despite high-profile pronouncements, experts describe the U.S. federal government’s AI oversight as disjointed and lacking cohesive direction. Institutions designed to coordinate national security policy on emerging technologies appear impaired, leaving AI governance reactive rather than proactive. Recent initiatives by the Trump administration, such as a voluntary self-regulation pledge signed by leading tech firms, have been criticized for their vagueness and limited enforcement mechanisms. The appointment of Director of National Intelligence Jay Clayton, who has limited AI expertise, as head of a newly formed “Super Intelligence Force” has raised further questions about the administration’s preparedness.
The geopolitical stakes surrounding AI are significant. Observers stress that the technology could establish a “recursive advantage” for whichever nation attains early leadership, reinforcing economic and military dominance over rivals. Former Biden administration officials suggest that while a broad bilateral accord on AI is unlikely, limited agreements on safety protocols and communication channels to manage dangerous incidents may be achievable. Treasury Secretary Bessent’s recent meeting with Chinese Vice Premier He Lifeng sought to establish such mechanisms, though the structure and oversight of these efforts remain unclear.
Experts caution that meaningful progress may depend on starting with incremental steps, akin to early arms control agreements during the Cold War. Unlike nuclear weapons, however, AI systems possess the capacity to generate instructions autonomously, complicating risk management. As the Shenzhen summit approaches, the international community faces growing pressure to balance rapid technological advancement with effective safeguards before the consequences of inaction become irreversible.