China opts out of blueprint on military AI use

Some 60 countries, including the United States, on Tuesday signed up to a  "blueprint for action" that governs the responsible use of artificial intelligence on the battlefield.

However, some 30 nations that sent a government representative to the South Korean summit, including China, did not back the document.

What are the guidelines about?

The guidelines said all applications of AI in the military sphere would be "ethical and human-centric."

The document examines what risk assessments must be made and the importance of human control. 

"Appropriate human involvement needs to be maintained in the development, deployment and use of AI in the military domain, including appropriate measures that relate to human judgment and control over the use of force," it said.

The Dutch government said the focus of the summit was an "action-oriented" set of guidelines, including discussions about real-world developments such as AI-enabled drones being rolled out by Ukraine.

Included in the details added to the latest document was the need to stop AI from being used to proliferate weapons of mass destruction by entities, including terrorist groups.

Ukraine was among the countries that signed the agreement, as well as prominent NATO members France, Germany and the United Kingdom.

Because of its invasion of Ukraine, Russia was not invited to the summit and so was not a signatory.

The Responsible AI in the Military Domain (REAIM) summit in Seoul is the second of its kind after one was held in the Dutch city of The Hague last year.

That gathering saw about 60 nations, including China, endorse a modest "call to action" without a legal commitment to take any action.

How realistic is a wider agreement?

"We are making further concrete steps," Netherlands Defense Minister Ruben Brekelmans told the Reuters news agency. "Last year ... was more about creating shared understanding, now we are getting more towards action."

"We also need to be realistic that we will never have the whole world on board," Brekelmans said, in response to the number of states that did not sign. 

"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.

"There are no principles and agreements yet," Brekelmans later posted on X, the platform formerly known as Twitter. "This is necessary to use AI responsibly. And to confront countries that break rules."

Giacomo Persi Paoli, head of Programme Security and Technology at the United Nations Institute for Disarmament Research (UNIDIR), said it might be counterproductive to rush ahead with rules upon which everyone might not agree.

"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."

Other initiatives on the same issue, include the US government's declaration on responsible use of AI in the military launched last year.

The Seoul summit was co-hosted by the Netherlands, Singapore, Kenya and the United Kingdom seeking to allow multi-stakeholder discussions are not dominated by a single nation or entity.