On September 22, 2026, Reuters exclusively reported, citing two sources familiar with the matter, that Chinese AI startup DeepSeek will brief the United Nations Security Council this week on the risks of artificial intelligence. According to Reuters, the 15-member Council plans to meet on Wednesday, September 23 local time, to discuss AI and international security.
To be clear: this is an exclusive report based on unnamed sources, not an official announcement.
A rare gathering
According to Reuters, OpenAI CEO Sam Altman plans to attend in person and brief the meeting, while senior representatives of Anthropic are expected to take part; DeepSeek and other Chinese AI companies, including Moonshot, have been invited to make statements, though DeepSeek founder Liang Wenfeng is not expected to attend. Reuters noted that neither company immediately responded to requests for comment, and that plans remain fluid.
If accurate, this would be a rare moment: frontier AI labs from China and the United States discussing AI risk in the same diplomatic setting. Reuters did not disclose the format of DeepSeek's briefing or the content of the statements — that awaits the official agenda.
Why now
According to Reuters, the meeting comes weeks after several AI industry leaders called for a coordinated slowdown in AI development, warning that systems could soon improve on their own and slip beyond human control. The Council first discussed AI risks in 2023: China warned AI should not become a "runaway horse"; the US warned against using AI to censor or repress. Last week, Secretary-General António Guterres again warned that AI risks cannot be ignored — at odds with the Trump administration, which argues existing safeguards are sufficient. Reuters also reported that Treasury Secretary Scott Bessent said on Monday the US and China had agreed to further AI safety talks, while Chinese state media dismissed slowdown calls as a "Cold War playbook" meant to preserve American tech dominance.
Why it matters
The value is not the meeting itself, but three converging elements: the venue — the Council maintains international peace and security, and its taking up AI risk means AI has formally entered the traditional security agenda; the lineup — according to Reuters, Altman attends in person, Anthropic sends senior staff, DeepSeek and Moonshot are invited, a rare scene of Chinese and American labs discussing governance at one table; the timing — Reuters noted that Chinese lab leaders rarely publicly discuss AI "deceiving humans, slipping beyond control, or causing catastrophic harm," and have stayed largely silent in recent weeks, so a DeepSeek briefing would be a rare public statement.
How to read it: keep expectations level
Reuters wrote "two sources familiar with the matter said"; the companies have not commented and no official agenda has been published. Until official confirmation, the guest list, the content, and even whether the meeting goes ahead remain fluid. The right way to read this is as a signpost worth watching, not a settled fact — whether the Council becomes a new stage for China–US AI governance dialogue depends on what actually happens on September 23, and no one can call that now.
