AI Leaders Urge UN to Control Technology, US Advocates Against Strict Controls

Major artificial intelligence executives urged the United Nations on Wednesday to establish controls over the rapidly advancing technology they are developing, citing potential risks to humanity. However, the United States expressed opposition to stringent global governance structures, advocating instead for continued development and shared benefits.

Dario Amodei, CEO of Anthropic, warned that "If managed poorly, I even believe AI could be a risk to humanity as a whole." Sam Altman, CEO of OpenAI, echoed these concerns, stating, "We could lose control of the future to AI." Both leaders testified before the U.N. Security Council, emphasizing the need for safeguards to prevent AI from becoming unmanageable.

They also called for efforts to prevent the concentration of AI power within a single company or nation, while acknowledging the technology's potential to improve lives. British Foreign Secretary Ed Miliband, present at the Security Council meeting, stated that the tech leaders' warnings about potential disaster "cannot be ignored." The U.K. plans to prioritize AI control in its G20 leadership next year, aiming for global standards to ensure safe development.

In contrast, White House Science Adviser Michael Kratsios argued that concerns about AI control should not halt its development or lead to new global governance frameworks. "We want you, our allies and partners, to share in the benefits of this technology," Kratsios said.

The United States and China are actively competing for AI dominance. Chinese U.N. Ambassador Fu Cong stressed that all countries should have the freedom to choose AI technologies and warned against the formation of exclusionary tech blocs. He advocated for the promotion of open-source models, aligning with China's approach to AI development.

World leaders and tech executives are grappling with fundamental questions about who should control AI, particularly in life-or-death decisions. U.N. Secretary-General Antonio Guterres had previously warned of "killer robots," a concern echoed by experts and European leaders who suggest such a reality is imminent or already present.

Amodei called for international agreement on preventing AI from being used to build biological weapons, establishing methods to verify AI model capabilities, and creating common standards for testing and addressing misuse. French Foreign Minister Jean-Noël Barrot, whose country convened the meeting, likened the current state of AI to the runaway computer HAL 9000 in "2001: A Space Odyssey" and the dawn of the atomic age, stressing the need for international regulation.

Other officials also voiced their concerns. Liberian UN Ambassador Lewis G. Brown II cautioned against "hesitation" leading to "tomorrow's catastrophe." Ukrainian President Volodymyr Zelenskyy highlighted the urgency of human control over machines, noting the potential for AI to make battlefield decisions and the need for peace before such a point is reached. He cited an incident where a Russian drone appeared to use AI for autonomous targeting, killing three Ukrainians.

Experts like Kentaro Toyama, a professor at the University of Michigan, confirmed the feasibility of building AI-powered autonomous weapons. Data scientist Rumman Chowdhury emphasized the need for concrete accountability, asking "Who authorized a machine to select and attack a target, under what constraints, and who answers for the result?"

The U.N. discussions coincided with Chinese President Xi Jinping's visit to Washington. Discussions between U.S. and Chinese officials, including Treasury Secretary Scott Bessent, have touched on a potential notification mechanism for AI incidents impacting national security. However, former President Donald Trump has expressed a preference for continued U.S. leadership in AI development, rejecting calls for control in the race with China.

Regional bodies in Europe and Asia, along with individual nations and U.S. states, are also developing their own AI policies, indicating a fragmented approach to global AI governance.