Top executives from leading artificial intelligence companies warned the United Nations Security Council on Sept. 23 that poorly managed AI could pose a ‘risk to humanity as a whole’, urging immediate international action to establish safeguards.
The executives—OpenAI CEO Sam Altman, Anthropic CEO Dario Amodei, and Hugging Face co-founder Clément Delangue—testified before the 15-member council during the UN General Assembly, emphasizing the need for global standards to prevent misuse, loss of control, and potential threats to international security. Their remarks followed growing concerns from researchers and policymakers about AI’s accelerating capabilities outpacing regulatory frameworks.
Immediate Action & Core Facts
1. AI leaders call for global AI governance
The CEOs urged the UN Security Council to adopt measures including:
- A ban on AI use in biological weapons development.
- Global standards to test AI models for risks of loss of control or misuse.
- A notification system for significant AI-related security incidents.
- Evaluation and verification systems to hold nations accountable.
Dario Amodei stated that Anthropic would “slow down as much as necessary” to ensure each AI release meets safety standards, while acknowledging that risk management requires “an industry-wide and global scale.” He warned that AI could evolve into a “country of geniuses in a data center” within two years if unchecked.
Sam Altman cautioned that AI’s rapid advancement could lead to a scenario where “humans lose control of the future”, adding that no single entity should wield the most powerful AI models to impose a worldview on others.
2. UN Security Council debates AI as a security threat
The briefing marked the first time AI executives addressed the UN Security Council, reflecting a shift in how governments view AI—not just as a commercial tool but as a potential threat to international peace and security. A European diplomat noted that the council’s focus includes whether algorithms could trigger geopolitical instability or military conflict.
Deeper Dive & Context
Global Calls for AI Safeguards
The executives’ testimony followed Amodei’s recent proposal to slow AI development, a stance publicly supported by Altman—a rare show of solidarity between rivals. Both companies have faced scrutiny over safety practices after industry researchers highlighted AI’s potential for catastrophic harm.
Yoshua Bengio, a Canadian AI pioneer and co-chair of a UN-backed scientific panel, warned that the dangers of advanced AI are “real and imminent.” He emphasized that the technology’s unpredictability could lead to uncontrollable escalation in global systems.
Diverging Political Responses
While AI leaders advocated for regulation, U.S. President Donald Trump rejected what he described as a “globalist scheme” to control AI during his UN General Assembly speech. Trump asserted that the U.S. would “not rein in” AI growth, calling it “bigger than the industrial revolution.” He framed AI as a tool for national advancement, contrasting with the executives’ warnings of existential risks.
China’s Role in AI Governance
Ahead of a meeting between Trump and Chinese leader Xi Jinping, the U.S. and China discussed establishing a notification system for AI-related national security threats. This reflects ongoing U.S.-China talks on AI governance, though Trump’s opposition to global oversight complicates multilateral efforts.
Key Questions Raised
The UN Security Council’s discussion centered on unresolved questions:
- Who should control life-and-death decisions—humans or machines?
- How can nations prevent AI from becoming too powerful to regulate?
- What safeguards are needed to prevent AI-driven military escalation?
Long-Term Implications
Diplomats and experts argue that without international coordination, AI could outpace human oversight, leading to scenarios where autonomous systems make decisions with irreversible consequences. The council’s involvement signals a recognition that AI governance is no longer a technical or corporate issue but a matter of global security.