This file photo, provided by the foreign ministry, shows an international military artificial intelligence conference in The Hague on Feb. 16, 2023. South Korea and the Netherlands co-hosted the first global Summit on Responsible Artificial Intelligence in the Military Domain, REAIM 2023. (Yonhap)
SEOUL, Sept. 24 (Korea Bizwire) — A global commission on the military use of artificial intelligence (AI) has called for legally binding agreements to ensure decisions to use nuclear weapons remain under human control, not AI systems, as part of responsible military AI integration.
The Responsible AI in the Military Domain (REAIM) Global Commission, a public-private initiative led by the Netherlands and South Korea, made the recommendation in its first guidance report on AI governance in the military domain, Seoul’s foreign ministry said.
Dutch Prime Minister Dick Schoof announced the outcome of the report during a high-level U.N. Security Council meeting on AI and international peace and security in New York on Wednesday (local time).
“Critical decisions about the use of nuclear weapons must unequivocally remain under human authority as they require moral, legal and strategic considerations,” the report said.
“These commitments should be pursued through either national policy declarations or an international agreement,” it said.
The “REAIM High-Level Meeting” is a Track 1.5 multilateral international forum launched to enhance global understanding of the military use of AI and its impact on international peace and security, and to contribute to the development of related international norms. (Image source: REAIM High-Level Meeting website capture)
The report came amid growing concerns over a “deficit,” or a lack of effective global governance, on military AI use and calls for the need to establish norms and guidelines for implementing safeguards, especially on decisions involving weapons of mass destruction, including nuclear weapons, the ministry said.
The commission was launched in 2023 on the initiative of South Korea and the Netherlands, both of which have hosted REAIM summits, to provide guidelines with experts’ participation.
The report recommended “anchoring” the development and use of AI in the military domain in relevant ethical principles and international law, to explain unlawful uses of AI and ensure the application of existing obligations, such as the right to use force and how force is employed during conflict.
It also called for implementing national policies that guarantee human responsibility across the AI system life cycle grounded in “rigorous test, evaluation, verification and validation.”
The guidance also calls for establishing a “permanent and inclusive” multilateral dialogue on the responsible integration of AI into the military, as well as developing an expert network to share knowledge and ensure all stakeholders are aware of AI governance frameworks.
(Yonhap)
