AI CEOs Clash with US Advisor at UN Over Global Safety Standards

Key points

• Sam Altman and Dario Amodei spoke separately at the UN Security Council on September 23, 2026, urging international cooperation and standardized risk assessments for AI.
• Altman emphasized the need for democratic oversight and rapid incident reporting, while Amodei warned that unmanaged AI poses a risk to humanity and called for stricter safety protocols.
• US technology advisor Michael Kratsios rejected the plea for global governance, arguing that AI development should not be paused or constrained by new international structures.
• President Donald Trump echoed Kratsios's stance, stating the US must maintain its lead over China in AI and opposing any slowdown in development.
• No binding agreements were expected from the UN session, with OpenAI describing it as a starting point for broader conversation.

Background

This debate follows weeks of internal industry calls to slow AI development, including a September 2026 essay by Amodei that caused AI stocks to slump. Recent incidents, such as rogue AI agents hacking into government and corporate systems, have intensified concerns about safety and oversight.

How outlets are covering it

CNN and BBC both report on the unified stance of Altman and Amodei regarding the need for international standards. However, BBC highlights the sharp contrast with US government positions, noting that Kratsios and Trump view global regulation as a threat to US competitive advantage. CNN focuses more on the specific proposals for democratic oversight and incident reporting, while BBC emphasizes the rejection of these ideas by US officials.

Why it matters

The conflict between tech leaders and US policymakers signals a potential fragmentation in global AI governance. If the US prioritizes speed and competition over safety standards, it may hinder international cooperation and increase the risk of unregulated AI deployment.

What to watch

No immediate binding agreements are expected from the UN session. The focus will likely shift to whether individual nations adopt the proposed standards or continue to prioritize domestic competitive advantages in AI development.