AI company leaders tell U.N. Security Council their systems could threaten humanity

# What happened

On Sept. 23, 2026, leaders of major artificial intelligence companies briefed the United Nations Security Council. Dario Amodei, CEO of Anthropic, and Sam Altman, CEO of OpenAI, testified that advanced AI systems could become a risk to humanity if mismanaged.

# Why the CEOs went to the U.N.

The immediate purpose was to put the issue before a global body that can influence international norms and potential regulatory efforts. The testimony highlighted two related policy questions: who should decide life-and-death outcomes when AI is involved, and how to prevent runaway or misused systems.

# Specific concerns raised

• Risk of loss of control: Both CEOs said AI systems might reach a level of capability that would make them difficult to contain once deployed.
• Concentration of power: They urged action to avoid AI control being centralized in a single company or nation, which could increase geopolitical instability.
• Life-and-death decision-making: The testimonies asked a core ethical and policy question: should humans or machines make decisions that affect survival? This was presented as a fundamental governance issue.
• Lethal autonomous weapons: U.N. Secretary-General Antonio Guterres warned of "killer robots," linking autonomous weapons to the broader security conversation about AI.

# What the testimony does and does not do

This briefing put industry leaders' concerns directly to the Security Council and publicized their call for international safeguards. It did not produce specific regulatory text, binding agreements, or a timeline for action. The CEOs requested that the Security Council consider controls but did not present a detailed enforcement mechanism in the testimony covered by the reporting.

# Immediate implications for policy and industry

• The U.N. discussion increases pressure on national governments to consider coordinated rules or standards for advanced AI systems.
• The framing of AI as a potential existential or global security risk could push the conversation beyond voluntary industry guidelines toward formal international measures.
• Calls to avoid concentration of power may influence antitrust, export control, and international cooperation debates.

# What to watch next

• Any follow-up by the companies or multilateral institutions proposing specific control mechanisms or verification processes.

# Bottom line

Top AI executives warned the U.N. Security Council that advanced AI systems could become a global hazard if left unchecked. They asked for international safeguards, highlighted the problem of concentrated power, and pressed the Security Council to address who decides on life-and-death outcomes when AI is involved. The briefings raised political and ethical questions that the U.N. and its member states must now consider further.