Anthropic Chief Executive Officer, Dario Amodei, has urged world leaders to work together against the growing risks of artificial intelligence. He warned that poorly managed AI could eventually threaten humanity.
Amodei spoke on Wednesday at a high-level United Nations Security Council meeting on AI and international security. The meeting was held on the sidelines of the 81st UN General Assembly in New York.
He addressed the 15-member council remotely and said the fast pace of AI advancement was creating both opportunities and serious security risks.
“If managed poorly, I even believe that AI could be a risk to humanity as a whole,” he warned.
The Anthropic CEO called on governments to work with technology companies to set common safety standards as AI systems become more capable.
He said Anthropic would “slow down as much as necessary” to make sure that each new AI technology the company releases is safe.
Amodei identified two broad categories of risk. The first is the deliberate misuse of AI, including its possible use in biological or cyber attacks. The second is the possibility that increasingly advanced systems could develop capabilities their creators cannot control.
He urged countries to begin with specific international agreements, including restrictions on dangerous applications such as using AI to develop biological weapons. He also called for systems to test and verify AI safety commitments.
“No leader, no company, and no nation can manage this alone,” Amodei told the council, stressing the need for international cooperation.
His appearance came alongside that of OpenAI CEO Sam Altman, who also warned that humanity could eventually lose control of AI if its development outpaces humans’ ability to understand or intervene in how it operates.
France convened the meeting, which was described as an effort to put AI safety and its implications for international peace and security at the centre of discussions among world powers.
