[Contribution] Towards responsible AI in military domain

On Sept. 9-10, South Korea will host the second summit on the Responsible Artificial Intelligence in the Military Domain (REAIM). The first-ever summit on this topic was held last year in the Netherlands.

REAIM 2024 in Seoul will continue to deepen understanding and awareness of the issues surrounding the application of AI in the military domain, and will explore ways to ensure that AI is used responsibly by militaries around the globe.

AI is transforming all facets of human existence at an astoundingly high pace. Ensuring that this technology is used responsibly is a challenging race against the clock. This holds especially true for the military application of AI.

While AI can be beneficial for the armed forces, the opposite is also true. On the one hand, commanders could rely on AI to quickly analyze complex data to help support effective decision making, potentially leading to more effective and precise operations and less collateral damage. On the other hand, however, improper understanding and oversight might make the situation worse. For instance, biases in training data and lack of human oversight and control could lead to poor operational judgments which could result in unnecessary destruction and suffering. It is therefore vital to advocate for the responsible application of military AI in accordance with international law. The breakneck speed with which AI develops and militaries seek to employ this ground-breaking technology requires us to act now.

REAIM 2024 builds on the discussions and outcomes of the first REAIM summit in 2023, which sought to increase political awareness regarding this topic as a means of ultimately moving towards the development of international agreements on the application of AI in the military domain. Stakeholder delegates from governments, businesses, academia and civil society of over eighty countries attended the summit.

Since the first REAIM summit, a Global Commission on REAIM has been set up, which is a body of preeminent AI scholars that seeks to support fundamental norm development and policy coherence across the globe related to military AI. Furthermore, several regional follow-up meetings were held as part of the REAIM process, in Asia, Africa, Latin America, Europe and North America, and the Middle East and Central Asia.

AI applied in the military domain is a cross-border, cross-discipline and multistakeholder phenomenon that can only effectively be tackled through the combined efforts of the global community. The REAIM process has proved vital in raising awareness, realizing mutual understanding and acquiring additional perspectives. Thanks to the efforts of our wide ranging global partners, we are now well positioned to explore norm development in the realm of AI in the military domain during the upcoming REAIM Summit.

South Korea – together with its co-hosts Singapore, Kenya, the United Kingdom and the Netherlands – is uniquely positioned to organize REAIM 2024. As a high-tech powerhouse, South Korea produces AI enabling HMB (High Bandwidth Memory) chips, possesses its own generative AI models, and produces innovative defense materials that leverage this technology. Moreover, as a Global Pivotal State, South Korea has led the global debate about AI in general by effectively hosting two additional AI-related summits in 2024. At REAIM 2024, the international community has another opportunity to ensure that the revolutionary power of AI is used responsibly and for the good of global security and stability.

Views expressed in this article are the writer's own. -- Ed.