China and the US are competing for AI dominance but have shared concerns over safety

Leaders from the United States and China are preparing for a high-stakes bilateral summit this week against a backdrop of escalating technological rivalry and deepening anxiety over the existential threats posed by rogue artificial intelligence. The impending discussions bring together U.S. President Donald Trump and Chinese leader Xi Jinping as their respective administrations vie for global supremacy in advanced computing, a race that President Trump has characterised with the blunt assessment that whoever wins AI, wins. Yet, even as both superpowers maintain fierce competitive postures and implement strict economic boundaries, diplomatic channels have opened to explore vital safety guardrails. As detailed in the Courant reporting, preliminary talks in New York between U.S. Treasury Secretary Scott Bessent and Chinese Vice Premier He Lifeng have laid the groundwork for potential crisis-management frameworks, highlighting an urgent diplomatic effort to mitigate shared technological perils without sacrificing national advantage.

The push for structured communication stems from a mutual recognition that uncontrolled model failures, biological misuse, and AI-enabled cyberattacks pose cross-border dangers that transcend geopolitical hostilities. Sun Chenghao, a senior fellow at the Center for International Security and Strategy at Tsinghua University, noted that there is still meaningful room for cooperation on risks that threaten both societies, observing that Beijing is paying increasing attention to loss-of-control scenarios and malicious threats. According to Eastbaytimes coverage, U.S. Treasury Secretary Scott Bessent told reporters that Washington wants to establish a shared vision of common goals and common threats through a proposed notification mechanism designed to handle emergencies that escalate to the national security level.

Media additions

Michael Kratsios, the science and technology adviser to the U.S. President, reinforced this perspective during a television appearance, emphasizing the necessity of avoiding shared risks through direct dialogue. Parallel insights from the Orange County Register reporting indicate that both governments harbour acute anxieties regarding the protection of vital services. In China, official governance frameworks specifically designate finance, electricity, telecommunications, and transportation as sectors uniquely exposed to frontier risks. Samm Sacks, a senior fellow at the Johns Hopkins School of Advanced International Studies, observed that Chinese experts have shown genuine willingness to engage with American counterparts on incident information sharing.

Despite these overlapping safety imperatives, the broader bilateral relationship remains heavily constrained by aggressive technological competition. President Trump has consistently downplayed appeals from domestic technology executives urging a slowdown in development for safety reasons, prioritizing American technological preeminence above all else. Conversely, Beijing has repeatedly condemned U.S. Export controls as an exclusionary tactic aimed at preserving an American monopoly by restricting access to elite computer chips. Alfredo Montufar-Helu of the consultancy Ankura suggested that any relaxation of these semiconductor and hardware restrictions is firmly off the agenda for the summit, leaving the practical room for substantive collaboration remarkably narrow.

The competitive friction has been further inflamed by recent developments in model performance and corporate accusations. As reported by the Taylorville Daily News, Chinese models engineered by firms such as Moonshot AI and DeepSeek have demonstrated capabilities that frequently rival those of American pioneers like OpenAI and Anthropic, often achieved at significantly lower operational costs. While U.S. Export restrictions have not halted China's progress, Will Cong, a professor at Nanyang Technological University, noted that they successfully throttle the rate of expansion by limiting available compute power. This computational squeeze has prompted Chinese developers to pivot increasingly toward domestic silicon, bolstering Beijing's confidence ahead of the summit, according to George Chen of The Asia Group.

Tensions have also flared over allegations of intellectual property extraction. U.S. Firms, most notably Anthropic, have accused Chinese competitors of utilizing a process known as distillation to siphon capabilities from frontier American models, effectively routing user requests to piggyback on U.S. Infrastructure. Rebecca Arcesati of the Mercator Institute for China Studies noted that Chinese labs are likely leveraging these methods to offset hardware limitations imposed by Western sanctions. Chinese officials and industry analysts have firmly rejected these accusations, arguing that distillation is a standard practice across the global software industry and does not account for the genuine innovative capacity of domestic models. Furthermore, Beijing has expressed its own security grievances, pointing to potential threats from American platforms such as Anthropic’s Claude systems and the risk that foreign tools could amplify political misinformation.

The geopolitical backdrop of the summit extends to broader economic and diplomatic alignments. Analysts tracking the current negotiations emphasize that technical confidence can only be built incrementally. Xiao Qian, vice dean of the Institute for AI International Governance at Tsinghua University, cautioned in the Telegraphherald coverage that meaningful safety cooperation remains impossible if either superpower reflexively classifies the other's technological advancement exclusively as a hostile national security threat.

Additional context from The Joplin Globe underscores that industrial exhibitions and global conferences earlier in the year highlighted China's aggressive domestic hardware push, showcasing liquid-cooled architectures and specialized accelerators designed to circumvent foreign embargoes. These domestic advancements ensure that President Xi approaches the negotiating table from a position of industrial resilience rather than dependency.

What to Watch Next

• The formal outcomes of the presidential meetings scheduled for this week, specifically regarding whether a binding incident-notification channel is established.
• Any subsequent statements from the U.S. Treasury and Chinese trade ministries concerning critical infrastructure protection protocols.
• Further developments in domestic Chinese semiconductor manufacturing and its impact on global frontier model parity.

As outlined across United News of Bangladesh reporting, the upcoming talks mark a critical juncture. Observers in both Washington and Beijing will be closely monitoring whether the two nations can successfully compartmentalize shared existential safety fears from their fierce, uncompromising race for technological supremacy.