Speakers including Dr. Radha Plumb, Chief Digital and Artificial Intelligence Officer, U.S. Department of Defense, sermon liable AI usage successful a plenary league astatine the Responsible AI successful the Military domain (REAIM) acme successful Seoul, South Korea, connected Sept. 9, 2024.
Reuters
About 60 countries including the United States endorsed a "blueprint for action" to govern responsible use of artificial quality successful the military on Tuesday, but China was among those which did not enactment the legally non-binding document.
The Responsible AI in the Military Domain acme successful Seoul, the 2nd of its kind, follows 1 held successful The Hague past year. At that time, astir 60 nations including China endorsed a humble "call to action" without legal commitment.
Government representatives said connected Tuesday said this year's "blueprint" was much action-oriented, successful keeping with precocious discussions and developments successful the military such as AI-enabled drones being rolled out by Ukraine, which also endorsed the document.
"We are making further factual steps," Netherlands Defense Minister Ruben Brekelmans told Reuters. "Last twelvemonth … was much about creating shared understanding, present we are getting much towards action."
This includes laying out what benignant of hazard assessments should beryllium made, important conditions specified arsenic quality control, and however confidence-building measures tin beryllium taken successful bid to negociate risks, helium said.
Among the details added successful the papers was the request to prevent AI from being used to proliferate weapons of wide demolition by actors including violent groups, and the value of maintaining quality power and engagement successful atomic weapons employment.
There are galore different initiatives connected the issue, specified arsenic the U.S. government's declaration connected responsible use of AI in the military launched past year.
The Seoul summit — co-hosted by the Netherlands, Singapore, Kenya and the United Kingdom — aims to guarantee multi-stakeholder discussions are not dominated by a azygous federation oregon entity.
However, China was among astir 30 nations that sent a authorities typical to the acme but did not backmost the document, illustrating stark differences of views among the stakeholders.
"We besides request to beryllium realistic that we volition ne'er person the full satellite connected board," Defense Minister Brekelmans said.
"How bash we woody with the information that not everyone is complying? … That is simply a analyzable dilemma that we should besides enactment connected the table," helium added.
The venue and timing for the adjacent acme is inactive being discussed, officials said.
At the UN General Assembly successful October, South Korean officials said they program to rise discussions about AI in the military domain based connected the 'blueprint'.
Giacomo Persi Paoli, caput of Program Security and Technology astatine the United Nations Institute for Disarmament Research, said countries should prosecute with others successful betwixt summits to mitigate immoderate risks.
"The blueprint is an incremental measurement forward," helium said. "By going excessively fast, excessively soon, determination is simply a precise precocious hazard that many countries do not privation to engage."
English (US) ·