DeepSeek to Brief UN Security Council as AI Risks Take Centre Stage

Chinese artificial intelligence startup DeepSeek is expected to brief the United Nations Security Council this week on the dangers posed by AI, as heads of state converge on New York for the annual General Assembly, according to two people familiar with the arrangements.

The 15-member Council has scheduled a Wednesday session on artificial intelligence and international security. OpenAI chief executive Sam Altman is also planning to address the meeting, where diplomats anticipate the presence of senior representatives from Anthropic.

DeepSeek and other Chinese AI developers, including Moonshot, have been invited to make statements, one source said. DeepSeek founder Liang Wenfeng is not expected to attend in person, the source added, noting that the schedule remains fluid and could shift at short notice. Neither DeepSeek nor Moonshot responded immediately to requests for comment.

The Council first examined the risks of artificial intelligence in 2023. During that discussion, China cautioned that the technology should not become a "runaway horse," while the United States warned against its use to censor or repress populations.

The upcoming meeting follows calls from several AI industry leaders for a coordinated slowdown in the development of increasingly capable systems, which they warn could soon improve themselves and escape human oversight. UN Secretary-General Antonio Guterres said last week that fast-moving AI advances carry risks that cannot be ignored — a position at odds with US President Donald Trump, who has argued that existing safeguards are adequate and that China would gain from doubts being cast on AI development.

Chinese state media have rejected calls for a development pause, describing them as a "Cold War playbook" intended to preserve Washington's lead in cutting-edge technology and keep Beijing out of global AI governance. Unlike their counterparts at US frontier laboratories, the heads of China's leading model developers have seldom publicly raised the prospect that advanced AI systems could deceive humans, evade control or cause catastrophic harm, and have largely stayed quiet as the debate intensified in the United States this month.

Separately, senior US and Chinese officials agreed to hold further talks on AI safety, US Treasury Secretary Scott Bessent said on Monday, after a meeting held ahead of Trump's scheduled discussions with Chinese President Xi Jinping later this week in Washington.