Public transcript

OpenAI CEO Sam Altman warns UN Security Council on AI risks

By: C-SPAN
Length: 9 min
Language: English
Text from: the video's captions
Transcribed: 13 October 2026

Read the full transcript Open on YouTube

Notes in English.

Summary

OpenAI CEO Sam Altman addresses the UN Security Council on the risks and opportunities of artificial intelligence. He warns of two major risks: losing control of the future to AI, and AI concentrating too much power in too few hands. He argues for international cooperation, common standards, and democratic oversight of AI development.

Key points

Altman says AI can either be a new renaissance of creativity and discovery or a new industrial revolution of upheaval and disarray.
He identifies two ways AI progress could go badly: losing control of the future to AI, and concentrating too much power in too few hands.
He states that no level of catastrophe risk, whether 10% or 0.1%, is remotely acceptable.
OpenAI has unilaterally slowed down in the past and will do so in the future rather than take excessive technological risk.
He cites rapid math progress, including a model solving the Navier-Stokes equations, one of the Millennium Prize problems.
He calls for complementary national and international frontier AI standards, incident reporting protocols, and secure information-sharing channels.
He insists the most important AI decisions cannot be made by labs in San Francisco alone and must be shaped through democratic processes.
Standards should not lock in incumbents and must support both open and closed model developers.

Questions it answers

00:00The choice before us

How does Altman frame the current moment in AI?

He says rapid model progress has compressed timelines, making both the upside and risks more immediate. AI can either be a new renaissance of creativity and discovery or a new industrial revolution of upheaval and disarray.

AI should give people more agency, not turn them into cogs in a machine.
There are many things AI cannot and should not automate.

02:02Potential and risks

What are the promises and dangers of more capable AI systems?

AI can help discover new knowledge, medicines, better healthcare, and stronger education, but as systems become more autonomous they could move faster than institutions or make decisions people no longer understand or control.

Recursive self-improvement could rapidly accelerate AI progress.
This moment calls for extreme care.

02:29Two ways things go wrong

What are the two major risks Altman warns about?

First, we could lose control of the future to AI as it moves too fast to follow or intervene. Second, AI could concentrate too much power in too few hands, letting one actor impose its worldview on everyone else.

No catastrophe risk level, from 10% to 0.1%, is remotely acceptable.
OpenAI has unilaterally slowed down in the past and will do so again.

04:41Three principles

What principles guide OpenAI's approach?

People must remain at the center of AI decision-making, the benefits of progress must be by and for people, and the technology must empower people individually.

Alignment is the work of keeping systems reliably under human control.
OpenAI sees itself as a builder, not the hero of the story.

05:59Math breakthroughs

What evidence does Altman give of AI's discovery capabilities?

Models went from okay at grade school math three summers ago to a gold medal level in the most prestigious international math competition, and recently solved the Navier-Stokes equations, a Millennium Prize problem.

The Navier-Stokes equations relate to aircraft design, weather forecasting, and blood flow.
The goal is empowering people, not solving math for its own sake.

06:59International standards

What international cooperation does Altman call for?

He calls for complementary national and international frontier AI standards, common ways to measure capabilities and assess risks, incident reporting protocols, and secure channels to share emerging vulnerabilities.

The most important decisions cannot be made by labs in San Francisco alone.
Standards must not lock in incumbents or favor one business model.

08:40Two paths forward

What are the two possible futures Altman describes?

One path concentrates power with opaque decisions where people feel AI is done to them; the other has democratic institutions guiding AI so it works for people and expands individual and community power.

Altman describes the second path as a future of realized potential, excitement, and wonder.
OpenAI is open to better ideas and ready to work with governments.

Notes

Framing the Moment

Sam Altman tells the UN Security Council that rapid model progress has compressed timelines, making both the upside and the risks more immediate.
He frames a choice: AI can be a new renaissance of creativity and discovery, or a new industrial revolution of upheaval and disarray.
The best version of AI gives people more agency, not optimizes the human parts of life away.

Two Major Risks

Losing control to AI: Systems could move faster than institutions, and recursive self-improvement could accelerate progress beyond human intervention.
Concentration of power: No person, company, or country should use the most powerful AI models to impose their worldview on everyone else.
Altman says no catastrophe risk level, whether 10%, 1%, or 0.1%, is remotely acceptable, and models should not be trained without a strong case for human control.
He notes OpenAI has unilaterally slowed down in the past and will do so again.

OpenAI's Three Principles

People must remain at the center of AI decision-making; alignment means keeping systems under human control.
Benefits of scientific progress and economic growth must be by people and for people.
The technology must empower people individually; OpenAI is a builder, not the hero of the story.

Evidence of Progress

Three summers ago models were okay at grade school math; last summer a model reached a gold medal level in the most prestigious international math competition.
A few weeks ago, a model solved the Navier-Stokes equations, a Millennium Prize problem used in aircraft design, weather forecasting, and blood flow study.

International Cooperation

The most important decisions cannot be made by labs in San Francisco alone; they must be shaped through democratic processes.
He calls for complementary national and international frontier AI standards covering capability measurement, risk assessment, and safeguards.
He urges accurate and speedy incident reporting and secure channels to share emerging vulnerabilities.
Standards must not lock in incumbents and should support open and closed model developers alike.