About 60 countries, including the United States, gathered at the Responsible AI in the Military Domain (REAIM) summit in Seoul to endorse a “blueprint for action” regulating the responsible use of artificial intelligence (AI) in the military. However, China was one of the countries that did not support the non-binding document, showcasing differing perspectives among stakeholders.
This year’s blueprint builds upon last year’s “call to action” with a more action-oriented approach. The document outlines risk assessments, human control conditions, and confidence-building measures to manage risks associated with AI in the military.
Key additions to the blueprint include preventing AI from being used for weapons proliferation and emphasizing human control in nuclear weapons deployment. The summit, co-hosted by the Netherlands, Singapore, Kenya, and the United Kingdom, aims to foster multi-stakeholder dialogues on AI in the military domain.
Despite some countries not endorsing the document, discussions on responsible AI use in the military will continue. The next summit’s venue and timing are still under discussion, with South Korea planning to address AI in the military at the UN General Assembly in October.
Giacomo Persi Paoli from UNIDIR emphasizes the need for countries to engage with each other between summits to mitigate risks, highlighting the importance of incremental progress in regulating AI in the military.
© Thomson Reuters 2024.