Sixty Countries Endorse ‘Blueprint’ for Military AI Use; China Decides Not to Participate, Reports Reuters

By Joyce Lee

SEOUL (Reuters) – On Tuesday, approximately 60 nations, including the United States, endorsed a "blueprint for action" aimed at establishing guidelines for the responsible use of artificial intelligence (AI) in military contexts. However, China was among the countries that chose not to support this non-binding document.

The Responsible AI in the Military Domain (REAIM) summit in Seoul marked the second such gathering, following one held in The Hague last year. At that earlier meeting, around 60 countries, including China, had endorsed a more limited "call to action" without legal obligations.

Representatives from various governments indicated that this year’s "blueprint" focuses more on actionable measures, in light of advancements in military technology, such as the deployment of AI-enabled drones by Ukraine, which also supported the document.

"We are making further concrete steps," said Netherlands Defence Minister Ruben Brekelmans. "Last year … was more about creating shared understanding, now we are getting more towards action."

The blueprint outlines necessary risk assessments, emphasizes the importance of human control, and discusses confidence-building measures to manage risks. Among its provisions, the document calls for preventing the use of AI in the proliferation of weapons of mass destruction, particularly by terrorist groups, and stresses the need for human oversight in nuclear weapons employment.

Various initiatives addressing this issue exist, including a declaration by the U.S. government regarding responsible AI use in the military, launched last year.

The Seoul summit, co-hosted by the Netherlands, Singapore, Kenya, and the United Kingdom, aims to ensure that discussions are inclusive and not dominated by a single nation or entity.

China did participate in the summit, sending a representative, but did not support the document, highlighting significant differences in perspectives among the participating countries.

"We also need to be realistic that we will never have the whole world on board," Minister Brekelmans noted. "How do we deal with the fact that not everyone is compliant? … That is a complicated dilemma that we should also put on the table."

Discussions regarding the next summit’s venue and timing are still ongoing. South Korean officials have expressed plans to address AI in the military domain at the upcoming UN General Assembly in October, based on the newly released blueprint.

Giacomo Persi Paoli, head of Programme Security and Technology at the United Nations Institute for Disarmament Research (UNIDIR), urged countries to engage with one another between summits to reduce risks. "The blueprint is an incremental step forward," he stated. "If we move too quickly, there is a significant risk that many nations may be hesitant to participate."