DeepSeek to Brief UN Security Council on AI Risks This Week, Reports Reuters
September 22, 2026 - 11:16 am
United Nations Security Council Chamber at United Nations Headquarters in New York City Credit: Mikaela Ryden via Shutterstock
According to an exclusive Reuters report cited by two sources familiar with the matter, DeepSeek is scheduled to address the UN Security Council this week regarding the risks associated with Artificial Intelligence (AI). This will take place during a meeting on Wednesday, coinciding with world leaders' attendance at the UN General Assembly.
DeepSeek, along with another Chinese AI company, Moonshot, has been invited to present, although the plans might still be subject to change, as per one of the sources. The source also reveals that invitations have reached several other Chinese AI firms, including Moonshot. It is noteworthy that DeepSeek's founder, Liang Wenfeng, may not attend personally.
American counterparts will also participate, with OpenAI CEO Sam Altman providing a brief in-person, as confirmed by an OpenAI spokesperson last week. Diplomats have informed Reuters that senior officials from Anthropic are expected to join as well. France, holding the rotating presidency of the council this month, is organizing the meeting.
This session follows recent debates on the rapid advancement of frontier AI. Dario Amodei from Anthropic suggested slowing down model improvements, a proposal supported by Altman and Elon Musk, but opposed by Mark Zuckerberg from Meta.
Governments remain divided on the issue: UN Secretary-General António Guterres emphasized the dangers of rapidly developing AI, while US President Donald Trump argued that existing safeguards are adequate.
The Security Council first discussed AI in 2023, when China expressed concerns about AI becoming a "runaway horse," and the United States raised issues related to its use for censorship and repression. A concept note from France for this week's session highlights potential malicious uses of AI and the risk of losing control over advanced models.
DeepSeek, based in Hangzhou, gained global attention in 2025 when its low-cost R1 reasoning model significantly impacted US technology stocks. This week's briefing will mark DeepSeek's participation in the same UN discussion as top US research laboratories.
By Ana-Maria Stanciuc
Editor-in-Chief, TNW
I am the Editor in Chief for TNW, covering technology not as a parade of launches and valuations, but as a system of influence, persuasion, and change. I write about startups, venture capital, digital policy, and Europe's ecosystem, focusing on the broader narrative beneath these trends: who shapes the future, who benefits from it, and how Europe is learning to assert its voice in global tech discussions. Prior to taking on senior editorial roles, I've spent over a decade in journalism, storytelling, content strategy, SEO, and digital marketing.