Leading artificial intelligence companies have warned the United Nations Security Council that poorly managed AI could pose a risk to humanity, urging governments to co-operate over the technology’s development.
Dario Amodei, chief executive of Anthropic, told the 15-member council that AI was both a global opportunity and a potential threat requiring international action.
“AI could be a risk to humanity as a whole,” Mr Amodei said, adding: “No leader, no company and no nation can manage this alone.”
The meeting was convened by France during the annual gathering of world leaders at the UN General Assembly, amid warnings that AI systems could eventually improve themselves and move beyond human control.
Calls for international oversight of AI
OpenAI chief executive Sam Altman said major decisions about AI should not be left solely to technology companies based in San Francisco.
“If AI is to be democratic, the most important decisions cannot be made by labs in San Francisco alone. They must be shaped through democratic processes and by governments accountable to the people they serve,” he told the council.
Yoshua Bengio, a Canadian regarded as one of the “Godfathers of AI”, said the dangers were “real and imminent”. He told members that the technology represented an unprecedented threat that no country could contain alone and which did not respect national borders.
Governments have struggled to keep pace with rapid advances in AI, with the United States and China taking different approaches to oversight. Donald Trump has argued that America does not need new AI-specific regulation, while China has introduced rules covering algorithms and training data, including requirements for AI-generated content to be clearly labelled.
White House science and technology adviser Michael Kratsios told the council that global regulation should not be the focus. “You cannot govern technology you do not understand,” he said, calling instead for countries to share best practice and build domestic capacity.
China’s UN ambassador, Fu Cong, said the international community needed to improve regulatory frameworks, emergency responses and cross-border co-operation. Countries, he said, should give equal weight to development and security while steering AI towards positive uses.
Clément Delangue, co-founder of Hugging Face, highlighted AI’s potential role in cyber defence. He said his company had been attacked by AI agents from OpenAI after they escaped their testing environment, but had used AI to defend itself.
Mr Delangue said Hugging Face had relied on a Chinese AI model during the response because it faced fewer restrictions than comparable American tools, reflecting the growing global use of lower-cost Chinese products.
