China declines to sign agreement on AI control of nuclear weapons

At the Responsible AI in the Military Domain (REAIM) summit in Seoul, which concluded Tuesday, nearly 100 participating countries, including the United States, China, and Ukraine, endorsed a non-binding declaration emphasizing the necessity for human oversight in decisions related to nuclear weapons.

The “Blueprint for Action,” adopted after two days of discussions, asserts that critical decisions regarding nuclear weapon deployment should remain under human control and involvement. The agreement also mandates that AI capabilities in the military domain must comply with national and international laws, and that AI applications should be ethical and human-centric.

China did not sign the agreement, and the Chinese embassy in Seoul did not provide immediate comment on the matter.

The declaration underscores that AI is currently employed for various military functions, such as reconnaissance, surveillance, and analysis, with the potential for future autonomous targeting. However, the summit did not specify any sanctions or penalties for violations of the agreement.

The summit, which excluded Russia due to its invasion of Ukraine, also recognized that further discussions are necessary to develop clear policies and procedures to keep up with the rapid advancement of AI in the military sector.

Co-hosted by Britain, the Netherlands, Singapore, and Kenya, the Seoul summit is presented as the most comprehensive and inclusive platform for discussing AI in the military domain, following the inaugural event in The Hague in February of the previous year.

Credit: The times of India