OpenAI, Anthropic Urge UN for AI Risk Standards

In September 2026, key figures in AI, including Sam Altman of OpenAI, Dario Amodei of Anthropic, and Clement Delangue of Hugging Face, addressed the United Nations. They called for the establishment of global standards to manage risks associated with AI development. Altman emphasized the need for common risk evaluation standards and international coordination to ensure AI technologies are developed safely. This call for action comes amidst differing opinions from political figures like Michael Kratsios, a former advisor to US President Donald Trump, who opposed slowing AI development due to competitive advantages over China.

Kratsios and Trump both highlighted the strategic edge the US holds in AI technology, arguing against any pause in development. Trump, in particular, emphasized maintaining the lead over China as a priority, underscoring the geopolitical stakes involved. These discussions occurred during a UN conference that took place earlier in September, marking a significant moment of debate on how AI should be governed internationally.

The push for international AI standards marks a potential shift in how AI governance is approached globally. If implemented, these standards could alter the competitive landscape, potentially leveling the playing field between nations. This could diminish the unilateral advantages currently seen in countries like the US. For AI companies, adhering to unified standards might streamline development processes but also impose new compliance costs.

The divergence in views between AI executives and US political figures highlights a growing tension between industry and government priorities. While tech leaders advocate for safety and collaboration, political leaders focus on maintaining strategic advantages. This dynamic could influence future policy decisions and shape the trajectory of global AI governance.

In the coming year, we can expect deliberations over these proposed standards to intensify. If the UN moves forward with Altman’s suggestions, it could lead to a framework that countries might adopt by mid-2027. This timeline would depend on consensus-building among major AI-developing nations, including the US, China, and EU member states.

The response from other global players will also be crucial. Countries like China may either align with these standards to foster international cooperation or resist them to maintain their competitive edge. Observing these dynamics will be critical to understanding how AI governance will evolve.

Should these standards be adopted, we might see shifts in AI investment patterns. Companies may prioritize compliance and safety enhancements over rapid innovation. This could benefit those already investing in robust risk management frameworks, such as European AI firms known for stringent data protection.

Additionally, the adoption of global standards could lead to increased regulatory scrutiny and enforcement, impacting companies that fail to meet these new benchmarks. This could prompt a reevaluation of AI development strategies, particularly for startups with limited resources.

In the broader context of sovereign AI, this development signifies a potential move towards shared governance, reminiscent of past international treaties on nuclear non-proliferation. However, unlike the Cold War era, the decentralized nature of AI development poses unique challenges. Countries will need to balance national interests with the collective benefits of global standards.

The tension between rapid technological advancement and the need for regulation is a recurring theme in AI policy. As global leaders navigate these waters, the outcome will likely influence AI's role in international relations for years to come.