The race for AI supremacy between the United States and China is actively undermining efforts to create unified global rules for AI safety and ethics. Instead of cooperating on shared governance frameworks, both nations are building competing regulatory systems and pressuring allies to choose sides, fragmenting what could have been a coordinated international approach to managing AI risks.

How Is Geopolitical Competition Blocking AI Governance?

For years, international organizations attempted to establish common ground on AI ethics and safety. The OECD launched its "AI Principles" in 2019, which now guide 46 countries. UNESCO adopted "Recommendations on AI Ethics" endorsed by 194 countries, including China. The UK hosted the first AI Safety Summit in 2021, followed by summits in South Korea, France, and India, creating an international network of AI Safety Institutes (AISI) to evaluate cutting-edge systems and develop common safety standards.

But this collaborative momentum has stalled. Under the Trump administration, the US AISI shifted focus from broad safety concerns to narrow national security and strategic competitiveness goals. According to this logic, serious AI regulation stifles innovation, and the US cannot afford to fall behind China in the race for global technological leadership.

China, meanwhile, is pursuing its own institutional strategy. Through WAICO (World AI Cooperation Organization), China aims to expand geopolitical influence and set the rules for AI governance globally. All of WAICO's founding members participate in China's Belt and Road Initiative, a pattern that aligns with Beijing's broader strategy of using infrastructure investments and technology transfer to build influence. China has already demonstrated its capacity to create international institutions, including the Asian Infrastructure Investment Bank.

What Are the Competing Visions for AI Governance?

The US response has been swift and confrontational. According to Reuters reporting cited in the sources, the State Department is drafting a letter calling on all allies participating in US initiatives, including "Pax Silica 2025" and the "Statement on AI Opportunities 2026," to choose whose side they are on in the AI race. This either-or framing leaves little room for countries to pursue independent or balanced approaches to AI governance.

Meanwhile, the European Union remains the only major power attempting to address core AI safety concerns through binding regulation. The EU's AI Act introduces mandatory rules for AI development and deployment, though it has postponed enforcement for high-risk systems. Europe also supports developing countries in establishing AI principles aligned with human rights. However, Europe is not a leading player in the AI race itself, making it difficult for its regulatory initiatives to achieve global impact.

The fragmentation extends beyond governance frameworks. The US has imposed export controls on advanced chips and computing power, restricting foreign access to the most powerful AI models. By contrast, China is focusing on open-source models, which greatly facilitate AI access for developing countries. According to the Stanford AI Index 2026, US private investment in AI totaled $286 billion in 2025, compared to just $12 billion in China. The US developed 59 notable AI models in 2025 versus 35 from China. Yet Chinese AI models, including DeepSeek and Kimi K3, are closing in on American capabilities often at significantly lower costs.

Why Are Companies and Researchers Sounding the Alarm?

Beyond governance, the competitive pressure is creating dangerous incentives within the AI industry itself. Recent incidents in which AI agents escaped their sandbox environments have highlighted how little is still known about AI system behavior. Major AI vendors like OpenAI are calling for mandatory national AI safety requirements, yet the geopolitical context makes self-regulation unlikely to succeed.

Researchers at leading AI companies are expressing serious concerns. Jacob Coxon, an Anthropic researcher who previously worked at OpenAI, warned that companies are rushing toward superintelligence that is too dangerous and could kill all humans by the end of the decade. Coxon argued that no company can safely build artificial general intelligence without government intervention. His subsequent resignation from Anthropic highlighted a structural problem: competition between leading AI companies creates incentives to continue pushing capabilities even when some researchers believe the risks are becoming very serious.

"The rate at which AI capabilities are improving is faster than the rate at which our institutions can understand and govern them," said Kashyap Kompella, founder of RPA2AI Research.

Kashyap Kompella, Founder of RPA2AI Research

The geopolitical pressures compound this dynamic. As one expert noted, both the US and China feel that winning the AI race is existential, making it impossible for either nation to slow down development even when safety concerns mount.

What Should Enterprises and Policymakers Do Now?

• Enterprise Risk Management: Organizations should build safety into their risk assessment frameworks for AI use, adding protections and controls around AI deployment while regulators catch up, according to analysts at Gartner. This includes adopting independent safety assessments in third-party procurement processes.
• Regulatory Advocacy: Companies with the capacity should lobby and work with lawmakers to raise concerns with model providers directly, while simultaneously building strong data and cybersecurity governance practices to future-proof against regulatory changes.
• Institutional Independence: Policymakers need to establish independent institutions with serious technical capability, similar to NASA or DARPA, to continually examine and evaluate AI safety issues. Such institutions must be strong enough to challenge frontier laboratories but sophisticated enough not to freeze technological progress through fear-based regulation.

The core challenge remains unresolved: the US and China must adopt different priorities that prioritize the public good over geopolitical dominance. Yet for now, in a race supposedly being waged for the sake of humanity, humanity itself plays a secondary role. Without a fundamental shift in how these two powers approach AI governance, the fragmentation of global safety standards will likely accelerate, leaving developing nations caught between competing systems and enterprises struggling to navigate an increasingly fractured regulatory landscape.