DeepSeek and US Firms to Brief UN Security Council on AI Risks

Chinese AI startup DeepSeek is expected to join a high-level briefing for the United Nations Security Council this week, addressing the emerging risks associated with artificial intelligence. According to two sources familiar with the matter, the session takes place as world leaders converge in New York for the annual General Assembly, marking a significant moment for international dialogue on technology governance.

The 15-member Security Council will convene on Wednesday to discuss the intersection of AI and international security. This meeting follows recent calls from industry leaders for a coordinated slowdown in the development of increasingly powerful systems, driven by concerns that such models could soon begin improving autonomously and potentially slip beyond human control.

Industry Leaders Prepare for Briefings

OpenAI CEO Sam Altman is also planning to address the council, with diplomats expecting senior representatives from Anthropic to participate in the discussions. DeepSeek and other Chinese AI firms, including Moonshot, have reportedly been invited to speak, though DeepSeek founder Liang Wenfeng is not currently expected to attend. One source cautioned that attendance plans could still shift at the last minute, while neither DeepSeek nor Moonshot immediately responded to requests for comment.

Divergent Views on Safety Measures

This is not the Security Council’s first examination of AI risks, as the body first took up the topic in 2023. At that time, China cautioned that AI should not become a “runaway horse,” while the United States warned against using the technology for censorship or repression. Recently, UN Secretary-General António Guterres warned that fast-advancing AI carries risks that cannot be ignored, a stance that contrasts with President Trump’s view that existing safeguards are sufficient.

Chinese state media have dismissed recent calls for a slowdown as a “Cold War playbook” intended to protect US dominance in advanced technology. Notably, unlike leaders at US frontier labs, heads of China’s top AI developers have remained largely quiet on the possibility that capable systems could deceive humans or cause catastrophic harm, despite the debate gaining momentum in the United States.

Bilateral Talks on AI Safety

Separately, US Treasury Secretary Scott Bessent stated on Monday that senior US and Chinese officials have agreed to hold further talks on AI safety. This development follows a meeting ahead of President Trump’s planned discussions with Chinese President Xi Jinping later this week in Washington, signaling a continued diplomatic engagement on the topic.