AI executives warn of risks and call for international safeguards at UN Security Council
The UN Security Council held a France-convened meeting on artificial intelligence and international security, hearing from leaders of OpenAI, Anthropic and Hugging Face. The executives warned that AI could outpace human oversight and called for stronger safeguards and international cooperation, while a senior US official opposed pausing development or creating new governance structures.
Key points
- The discussion combined warnings about AI risks with proposals for standards, incident reporting and wider public oversight.
Evidence · 5 source statements
He called for "national and international" AI standards regarding ways to measure capabilities of an AI tool, assessing related risks, AI safeguards, and the degree to which…
speedy incident reporting, classification and reporting protocols so the world can learn from failures before they become catastrophes.
- The speakers did not offer a single approach: Anthropic pledged to slow releases if needed for safety, while Hugging Face’s chief argued for open-source AI as a means of defense.
Evidence · 2 source statements
We will slow down as much as necessary in order to make sure that every successive AI technology that we release is actually safe,
The world needs open-source AI more than ever to defend itself.
- The meeting also exposed a policy divide between calls for international standards and the US administration’s opposition to new governance structures.
Evidence · 3 source statements
OpenAI has urged the United States to lead a global effort on establishing standards for AI, noting that its industry is at the technical frontier.
Michael Kratsios, director of the White House Office of Science and Technology Policy, said rapid advancement is not a reason to pause AI's development or constrain it with new…
He called for "national and international" AI standards regarding ways to measure capabilities of an AI tool, assessing related risks, AI safeguards, and the degree to which…
Who's involved
Sam Altman
OpenAI’s chief executive, he spoke at the meeting and proposed safeguards and a role for democratic institutions.
Altman called for national and international standards, rapid incident-reporting protocols and decisions shaped by democratic processes rather than by AI labs alone. He also said models should not be trained without strong evidence that people can keep them under control.
What they said · 4 source statements
If AI is to be democratic, the most important decisions cannot be made by labs in San Francisco alone,
He called for "national and international" AI standards regarding ways to measure capabilities of an AI tool, assessing related risks, AI safeguards, and the degree to which…
speedy incident reporting, classification and reporting protocols so the world can learn from failures before they become catastrophes.
Dario Amodei
Anthropic’s chief executive, he spoke at the meeting about AI safety and his company’s development practices.
Amodei warned that poorly managed AI could threaten humanity and said Anthropic would slow development as needed to make its successive releases safe.
What they said · 2 source statements
If managed poorly, I even believe that AI could be a risk to humanity as a whole,
We will slow down as much as necessary in order to make sure that every successive AI technology that we release is actually safe,
Clément Delangue
Hugging Face’s chief executive, he was among the technology leaders briefing the Council.
Delangue argued that the world needs open-source AI for defense and stronger standards for monitoring and disclosing incidents.
What they said · 2 source statements
The world needs open-source AI more than ever to defend itself.
The global community needs stronger standards for monitoring and incident disclosure.
OpenAI
The company’s chief executive spoke at the meeting, and the company has called for a US-led international standards effort.
OpenAI urged the United States to lead a global effort to establish AI standards.
What they said · 1 source statement
OpenAI has urged the United States to lead a global effort on establishing standards for AI, noting that its industry is at the technical frontier.
Michael Kratsios
As director of the White House Office of Science and Technology Policy, he represented the US administration’s approach to AI governance.
Kratsios said rapid AI progress did not justify pausing development or constraining it with new governance structures.
What they said · 1 source statement
Michael Kratsios, director of the White House Office of Science and Technology Policy, said rapid advancement is not a reason to pause AI's development or constrain it with new…
Yoshua Bengio
He is a member of the Independent International Scientific Panel on AI and offered an assessment of the risks.
Bengio said AI dangers were real and imminent, and criticized companies and countries for taking reckless risks in pursuit of self-interest.
What they said · 2 source statements
“The dangers are real and imminent.”
Today, a handful of companies and countries are acting in their self-interest by taking reckless risks with humanity's fate.
Chen Jing
The technology-policy researcher commented on international AI governance and the interests of developing countries.
Chen said AI risks cross national borders and called for cooperation, including UN-framework infrastructure for reporting vulnerabilities and detecting deepfakes, with technical assistance for the Global South.
What they said · 2 source statements
From the perspective of global AI governance, the participation of Chinese AI firms in the UN discussion highlights a key reality: AI-related risks now go beyond the governance…
For many countries in the Global South, AI is not only a security issue but also a development opportunity, Chen said, adding that under the UN framework, public infrastructure…
Liu Gang
The AI-policy economist addressed who should participate in international governance.
Liu argued that governance should include major technology players from China, the United States, Europe and the Global South, and should be inclusive, non-discriminatory and development-oriented.
What they said · 2 source statements
With Chinese firms now invited to participate, the debate may reflect a broader recognition that meaningful AI governance requires participation from all major technology…
For the international community as a whole, the core message is that AI governance should be inclusive, non-discriminatory and development-oriented, Liu added.
Where accounts differ
Whether AI development needs new governance structures and international standards
- OpenAI and Altman called for a US-led global standards effort and national and international standards for measuring capabilities, assessing risks and maintaining human oversight.
Evidence · 2 source statements
OpenAI has urged the United States to lead a global effort on establishing standards for AI, noting that its industry is at the technical frontier.
He called for "national and international" AI standards regarding ways to measure capabilities of an AI tool, assessing related risks, AI safeguards, and the degree to which…
- White House official Michael Kratsios said rapid progress was not a reason to pause AI development or constrain it with new governance structures.
Evidence · 1 source statement
Michael Kratsios, director of the White House Office of Science and Technology Policy, said rapid advancement is not a reason to pause AI's development or constrain it with new…
The statements reflect different approaches to the policy response: international standards and oversight on one side, and opposition to new governance constraints on the other.
What was reported
- France convened a Security Council meeting on AI and international security under the Council agenda item on maintaining international peace and security.
Reported independently by 2: France 24, Global Times
Evidence · 2 source statements
France convenes a UN Security Council meeting on Wednesday on artificial intelligence (AI) and international security that includes briefings by leading industry figures,…
The meeting, chaired by France, is listed by the UN website as a discussion on AI and international security, under the agenda item "Maintenance of international peace and…
- Executives from OpenAI, Anthropic and Hugging Face briefed the Security Council; the reported participants included Sam Altman, Dario Amodei and Clément Delangue.
Reported independently by 2: CNBC, Reuters
Evidence · 2 source statements
Executives from three of the world’s leading artificial intelligence companies briefed the UN Security Council on Wednesday amid warnings that increasingly powerful AI systems…
Other participants are OpenAI CEO Sam Altman, Anthropic CEO Dario Amodei and Hugging Face CEO Clément Delangue, CNBC reported.
- Altman warned that increasingly powerful and autonomous AI systems could develop faster than institutions can respond and make human monitoring or intervention harder.
One source: RIA Novosti
Evidence · 3 source statements
- Altman proposed common standards for assessing AI capabilities and compliance, along with systems for reporting incidents and secure communication among governments, infrastructure operators and technical experts.
One source: Yonhap News Agency
Evidence · 1 source statement
- Amodei said stronger safety standards would be needed as AI models grow more powerful and pledged that Anthropic would slow development as much as necessary to ensure its releases are safe.
One source: France 24
Evidence · 2 source statements
even more stringent standards of safety will be required.
We will slow down as much as necessary in order to make sure that every successive AI technology that we release is actually safe,
How coverage differed
Chinese state media — Global Times highlighted the expected participation of Chinese firms and Chinese experts’ calls for inclusive governance and attention to Global South development.
Reports emphasizing executive safety warnings — The Guardian and France 24 focused on warnings about AI risks, calls for caution and Anthropic’s pledge to slow development if necessary.
Still unclear
- No Security Council resolution, binding rules or formal agreement were announced.
Not heard from
- DeepSeek and Moonshot — Chinese firms were reported as expected to brief the Council, but their own statements or views are not included.
- France and other Security Council members — France convened the meeting, but the accounts here do not include statements from the French government or other Council members about the meeting’s outcome.