A comprehensive framework for the ethical use of artificial intelligence in military operations, including nuclear systems, was adopted by 61 countries at the Responsible AI in the Military Domain (REAIM) summit in Seoul.
The summit concluded with the adoption of a 'Blueprint for Action,' which was endorsed by 1,952 participants from 96 countries, including 38 ministerial-level officials.
The blueprint warns that military AI applications pose humanitarian, legal, security, and ethical challenges that must be addressed. It emphasizes the need to prevent AI from contributing to the proliferation of weapons of mass destruction.
It also states that AI technologies should support disarmament and arms control efforts, ensuring human oversight in critical decisions regarding nuclear weapons.
The document outlines 20 clauses focused on the impact of AI on international peace, responsible implementation in military contexts, and future governance.
Countries endorsing the blueprint include South Korea, the US, Japan, France, and the UK. Notably, China attended but did not support the framework.
Kenya's Defense Minister announced the country would join a core group leading the first UN resolution on responsible military AI, highlighting a global commitment to ethical AI in warfare.
South Korea's President Yoon Suk Yeol reaffirmed the country's leadership in global AI governance, further solidifying its role in establishing international norms for military AI.