OpenAI has proposed that the United States take the lead in establishing international standards for advanced artificial intelligence, focusing on incident reporting, cooperation frameworks, and safe development practices.
- OpenAI proposes US leadership for global AI safety and reporting standards.
- Calls for coordination via national AI safety institutes like the US Center for AI Standards and Innovation.
- Recommends halting fully autonomous self-improving AI until proven safe.
What happened
OpenAI publicly requested that the United States take charge of forming an international framework for advanced AI standards. Their focus includes establishing consistent rules for which AI incidents must be reported and defining protocols for cross-country cooperation and access to computational resources. OpenAI emphasized following the example set by the European Union’s AI Act, specifically Article 55, which mandates providers of high-risk AI models to report serious incidents promptly.
The request was accompanied by a suggestion to route these efforts through existing national AI safety entities, such as the US Center for AI Standards and Innovation housed within the Commerce Department. OpenAI also highlighted the necessity of not pursuing fully autonomous recursive self-improvement of AI models unless it can be demonstrated to occur safely. This comes amid rising awareness of the risks posed by AI systems capable of rapid autonomous advancement.
Why it matters
The push for US leadership in global AI standards underscores the importance of harmonizing regulations as AI technologies rapidly evolve. Currently, Europe has advanced governance structures in place to oversee AI safety and incident management, including a scientific advisory panel drawn largely from EU and EFTA countries. By contrast, the US lacks a permanent director at its primary AI standards institute and has no comparable regulatory body, leaving a leadership gap in AI governance.
Establishing consistent international standards is critical to manage systemic risks posed by increasingly powerful AI systems, including potential national security concerns. OpenAI’s proactive filing of incident reports under the EU AI Act demonstrates a commitment to transparency, while urging the US to codify similar mandatory rules for testing, cybersecurity, and evaluation reflects industry recognition of the need for government engagement and oversight.
What to watch next
Key developments to monitor include how the US government responds to OpenAI’s call for leadership and whether it builds out institutional capabilities comparable to the EU’s AI Office and scientific panels. The leadership vacancy at the US Center for AI Standards and Innovation and the recent departure of high-profile AI coordinators highlight organizational instability that could impact progress.
Additionally, upcoming international engagements, including OpenAI CEO Sam Altman’s address to the UN Security Council and his participation in diplomatic events, may signal increased coordination on AI governance. The continuation and expansion of mandatory AI incident reporting rules, as well as efforts to regulate autonomous self-improving AI, will be pivotal in shaping global AI safety frameworks.