UN Launches First Forum on Military AI Risks
The UN launches its first global forum on military AI on October 22, aiming to keep humans in control of weapons and set safety rules for the arms industry.
The United Nations is opening its first-ever international forum on the military uses of artificial intelligence on Thursday, October 22. The forum will bring governments, tech companies, and civil society together to deal with the growing risk that machines could make life-and-death decisions on their own.
A forum that opens today
The launch event takes place at UN Headquarters in New York at 13:00 EDT in Conference Room 4. It is organized by the UN Institute for Disarmament Research (UNIDIR) under its Centre of Excellence on AI, Peace and Security, in cooperation with Switzerland and Pakistan. According to the UN, this is the first and only multi-stakeholder forum inside the UN system that works on AI in international peace and security and in the military domain. The event will also open expressions of interest for organizations that want to take part.
Why human control is slipping
UNIDIR Director Robin Geiss set out the urgency at a UN press conference on Tuesday. He warned that "the margin for human judgment is increasingly narrow, including in the nuclear domain," adding that it is hard to tell how close the world is to machines acting beyond human control. "Artificial intelligence autonomy is already embedded widely across the military domain," Geiss said. "We find them in intelligence analysis and target identification, drones, cyber attacks, information operations, and so on as we speak." Decision cycles are compressing to machine speed, Geiss warned, and opaque systems raise the risk of miscalculation and unintended escalation. The falling cost of powerful AI tools also means non-state actors can now run sophisticated operations that once needed a whole government behind them. He noted that just last month, the United States proposed an AI safety hotline with China in case of an AI incident that could threaten national security.
Three goals: rules, definitions, and armed groups
The new forum will start work on three priorities. First, it will develop voluntary guidelines for responsible industry behavior, grounded in international law and built with the Office of the UN High Commissioner for Human Rights, to guide how companies design AI tools for military use. Second, it will build an AI security lexicon that clarifies key terms such as autonomy, safety, and security, which different countries and companies currently use very differently. An event on the lexicon is already scheduled for Friday, October 23. Third, together with the UN training arm UNITAR, the forum will examine how non-state armed groups, including terrorist organizations and criminal networks, could misuse AI for planning, logistics, and attacks - and what national and international measures could stop them without blocking the peaceful uses of AI.
What happens next
The forum gives the UN a permanent table where governments, defense tech companies, software developers, researchers, and civil society can talk about military AI openly. Geiss acknowledged that geopolitical competition is a serious obstacle, but said concerns about AI's impact are already crossing geopolitical lines and prompting action from major powers. The practical tests come fast: guidelines must be written with an industry that moves quickly, the lexicon must turn disputed words into shared definitions, and the misuse work must produce rules that governments can actually enforce.
Sources
UNIDIR - https://unidir.org/event/unidirs-centre-of-excellence-on-ai-peace-and-security-forum-launch/ UN Media (UNifeed) - https://media.un.org/unifeed/en/asset/d366/d3660366 Associated Press (via ETV Bharat) - https://www.etvbharat.com/en/technology/new-un-initiative-will-aim-to-tackle-the-military-uses-of-ai-enn26102100404 Dig.watch - https://dig.watch/updates/unidir-director-warns-of-military-ai-operations