DeepSeek and Dario Amodei to Advise UN on AI Risks

DeepSeek, a Chinese artificial‑intelligence startup, will address the United Nations Security Council on Thursday, joining former OpenAI researcher Dario Amodei on a panel that will focus on the security risks posed by rapidly advancing AI systems.

The meeting comes as governments and multilateral bodies scramble to keep pace with technology that could reshape warfare, disinformation and economic power. The UN began formal discussions on AI in 2023 after a series of high‑profile incidents, including deep‑fake videos used in elections and autonomous weapons trials in Russia and the United States. A 2023 UN report warned that unchecked AI could destabilise international peace, prompting member states to seek expert input from the private sector.

DeepSeek’s invitation is notable because the company is backed by Beijing’s leading venture capital firms and has released a large‑scale language model that rivals those of Western firms. Amodei, co‑founder of Anthropic and a vocal critic of China’s AI strategy, has spent the past year warning that the Chinese government’s lack of transparency makes it the “single biggest threat” to global AI safety. In response, U.S. and European delegations have expressed cautious optimism, saying they hope the dialogue will surface concrete mitigation measures. Analysts note that the combined presence of a Chinese firm and an outspoken Western expert underscores the UN’s intent to hear all sides before drafting policy.

The Security Council is expected to use the briefings to shape a new advisory group on emerging technologies, a step that could lead to the first binding UN resolution on AI governance. If the council reaches consensus, member states may be required to report AI development activities and adopt safeguards against weaponisation. The outcome could set a precedent for how the world collectively manages a technology that is already influencing economies, militaries and societies worldwide.