China refuses to sign agreement on AI control of nuclear weapons

A global summit on AI in the military domain concluded Tuesday with a non-binding declaration urging nations to maintain human control over nuclear weapons. Article summary.

While nearly 100 countries, including the United States, China, and Ukraine, participated in the Responsible AI in the Military Domain (REAIM) summit in Seoul, China notably refused to sign the agreement.

The two-day summit, co-hosted by Britain, Singapore, the Netherlands, and Kenya, resulted in a global call to maintain human involvement in all nuclear-related decisions.

The "blueprint for action" was endorsed by nearly 60 countries. It outlines key principles such as human control, risk assessment, and preventing AI from being used to proliferate weapons of mass destruction.

The guidelines highlight that AI in military contexts must adhere to national and international laws, with a strong ethical and human-centric approach.

While the summit was a follow-up to a similar event held in Amsterdam last year, the Seoul gathering focused more on concrete actions and addressed recent developments, such as Ukraine's deployment of AI-enabled drones.

Despite the consensus from most nations, China's decision not to sign the declaration sparked concerns over the future role of AI in warfare, particularly in the nuclear realm.

The advancement of digital technology has revolutionised how people go about their daily lives, and it's no surprise that these same tools could enhance military capabilities. However, military experts have warned about the potential dangers of AI-controlled weapons, particularly nuclear ones.

While AI is already used for reconnaissance, surveillance, and analysis, autonomous target selection remains a contentious issue.

The Seoul summit did not outline specific sanctions or punishments for countries that violate the agreement. Instead, the declaration acknowledged the need for further discussions to establish clear policies and procedures regarding AI in the military domain.

"We are making further concrete steps," Netherlands Defence Minister Ruben Brekelmans told Reuters.

"Last year...was more about creating shared understanding, now we are getting more towards action."

When asked about the China's refusal to sign the agreement, Brekelmans said we "need to be realistic that we will never have the whole world on board."

"How do we deal with the fact that not everyone is complying? ...That is a complicated dilemma that we should also put on the table," he added.

Giacomo Persi Paoli, head of Programme Security and Technology at the United Nations Institute for Disarmament Research (UNIDIR), suggested that rushing to establish rules without broad consensus could be counterproductive.

He argued that an incremental approach would be more effective in engaging a wider range of countries.

"The blueprint is an incremental step forward," he said.

"By going too fast, too soon, there is a very high risk that many countries do not want to engage."

Russia, also absent from the summit due to its ongoing invasion of Ukraine, has also not expressed formal views on the matter, leaving its stance on AI in warfare ambiguous.