Archive · 23 Sep 2026
OpenAI Urges US Leadership on Global Frontier AI Standards
OpenAI wants the United States to help countries create shared rules for very powerful artificial intelligence.
These rules would cover AI that might improve its own abilities.
OpenAI says countries should use common tests and report serious AI problems in similar ways.
The company warns that different national rules could make international risks harder to manage.
It says the United States is well positioned to lead because of its role in technology and other global networks.
OpenAI also says fully independent self-improving AI is not happening today.
It believes such systems should not be developed without human control and safety measures.
Sam Altman will discuss AI and international security with the United Nations Security Council.
The Trump administration does not support slowing AI development because it says a pause could help China catch up.
OpenAI called for the United States to lead global technical standards for frontier artificial intelligence, including recursive self-improvement.
The company proposed shared national and international standards, common measurements, and incident-reporting protocols.
OpenAI warned that differing national rules could fragment evaluations and make cross-border AI risks harder to address.
The company said fully autonomous recursive self-improvement is not happening today and should not proceed without human oversight.
Sam Altman is due to brief the 15-member United Nations Security Council, while the Trump administration opposes slowing AI development.
- Who
- OpenAI, chief executive Sam Altman, United States officials, other AI industry leaders, and United Nations officials.
- What
- OpenAI called for a United States-led international effort to establish technical standards for frontier AI, including recursive self-improvement.
- When
- The proposal was published on Monday, September 23; the United Nations Security Council briefing is scheduled for Wednesday.
- Where
- OpenAI published the proposal in a blog post as world leaders gathered at the United Nations; the Security Council meeting will be held at the United Nations.
- Why
- OpenAI said shared standards, measurements, and incident-reporting procedures are needed to manage cross-border AI risks and preserve meaningful human oversight.
Prioritize AI Safety and Pacing
Continue AI Development
Development speed
Prioritize AI Safety and Pacing
OpenAI and several AI industry leaders support pacing development so alignment research and its deployment remain ahead of AI capabilities.
Continue AI Development
The Trump administration opposes slowing development, arguing that a pause could help China close the gap in a technology considered important to national and economic security.
Global standards
Prioritize AI Safety and Pacing
OpenAI says the United States should lead shared international standards, measurements, and incident-reporting procedures to address risks that cross borders.
Continue AI Development
OpenAI warns that failing to establish leadership could result in a fragmented, uneven, and conflict-ridden global system, although the articles do not describe a specific opposing standards proposal.
Recursive self-improvement
Prioritize AI Safety and Pacing
OpenAI warns that autonomous self-improvement could outpace human understanding and oversight, potentially making AI more dangerous or less aligned.
Continue AI Development
The articles report arguments for continuing AI development but do not identify a specific group advocating unrestricted autonomous recursive self-improvement; OpenAI says it should proceed only if safely controlled.
- Proponent
- OpenAI
- Requested US role
- Lead international efforts to develop global technical standards for frontier AI.
- Technology covered
- Frontier AI, including recursive self-improvement (RSI).
- Proposed measures
- Complementary national and international standards, common measurements, and incident-reporting protocols.
- OpenAI safety position
- Fully autonomous recursive self-improvement is not happening today and should not be pursued unless human control and oversight can be preserved.
- Upcoming briefing
- Sam Altan is due to brief the 15-member United Nations Security Council on AI and international security on Wednesday.
- US administration position
- Donald Trump's administration rejects calls to slow AI development, saying a pause could give China room to close the technology gap.
