AI Leaders Urge Global Cooperation at UN, Warn of Existential Risks

The leaders of artificial intelligence companies OpenAI and Anthropic have urged the United Nations Security Council to establish global cooperation on AI safety standards, warning that the technology is at a critical juncture.

OpenAI CEO Sam Altman told the council that AI presents a choice between a "renaissance of creativity and discovery" or an "industrial revolution of upheaval and disarray." He emphasized that the decisions shaping AI's future cannot be solely made by a few companies in San Francisco, but must involve democratic processes and accountable governments.

Anthropic CEO Dario Amodei echoed these concerns, stating that AI, if poorly managed, could pose a risk to humanity. He highlighted Anthropic's commitment to prioritizing safety, even if it means slowing down the release of new models, and stressed that managing AI risks is a global endeavor beyond any single company.

Both executives acknowledged their companies' commitment to pausing development if significant risks are identified, though recent releases of new models by both firms have occurred. They called for international AI standards, robust incident reporting, and secure channels for sharing safety information.

Their calls for global regulation contrast with U.S. President Donald Trump's stance. Trump, in his address to the UN, stated the U.S. "rejects" a global attempt to regulate AI, asserting that the nation is leading in AI development and will maintain that position without stifling growth.

Altman also outlined a "middle path" for AI, cautioning against both excessive optimism and extreme pessimism. He cited AI's rapid advancements in mathematics, noting that models have progressed from performing at a middle school level three years ago to winning a gold medal at a prestigious international math competition and solving complex equations like the Navier-Stokes equations.

Clem Delangue, CEO of Hugging Face, shared an experience where his company was attacked by AI but defended itself using similar technology. He argued that the primary risk is not powerful AI itself, but the asymmetry of that power between different actors and nations.

Yoshua Bengio, a computer scientist and AI expert, provided opening remarks, pointing to recent AI actions that have alarmed the public and industry. He identified three global threats from advanced AI: catastrophic misuse, power concentration, and loss of control. Bengio urged collective agreement that no single entity should develop an AI system capable of causing large-scale harm.