New UN initiative will aim to tackle the military uses of AI

The United Nations is launching a new initiative focused on the military uses of artificial intelligence. The initiative, by the U.N. Institute for Disarmament Research, aims to address dire concerns about the potential threat AI could pose to humanity. The institute plans to establish the first international forum for key players to tackle this issue.

Robin Geiss, head of UNIDIR, warned at a U.N. press conference that the margin for human judgment is increasingly narrow, especially in the nuclear domain. The rapid development of AI makes it difficult to determine when machines may act outside of human control, and addressing this uncertainty is the main goal of the initiative.

AI is increasingly being used in various areas such as early warning systems, information gathering, and reconnaissance, making it challenging to pinpoint when control is lost. Geiss emphasized that while global competition to lead on AI poses obstacles, concerns about AI's impact are transcending geopolitical blocs and prompting action from major powers.

A proposal from the U.S. and China last month for an AI safety alert system to address issues threatening national security is one example. The U.N. recognizes that the risks posed by AI are a major topic at their annual high-level meeting, with leaders of many member nations discussing the issue. UNIDIR has established a Center of Excellence on AI, Peace and Security, which received funding from all five major nuclear powers.

This week, UNIDIR will launch a new forum as part of the center to bring together actors from every relevant sector and around the world to work on AI in military uses. The forum will be part of other AI efforts being launched by the center, including one focused on the role of industry and developing a framework for responsible industry behavior to ensure AI-enabled military capabilities are reliable, predictable, and follow international law.

UNIDIR is also launching an initiative to address the risks of AI misuse by armed groups, which are already using the technology for operational planning, execution, and logistics.