A Coruna, Spain: Limited agreement reached at military AI summit
A CORUNA, Spain, Feb 5 (Reuters) – In a notable development at the Military Artificial Intelligence Summit, about a third of participating countries reached agreement on a declaration aimed at governing the application of AI in war.
However, there was a big gap in this agreement, as the military powers China and the United States did not participate.
As several attendees and delegates noted, prevailing tensions between the United States and its European allies, along with uncertainty regarding the future of transatlantic relations, created hesitation among some countries regarding entering into a solidarity agreement.
The pledge highlights growing concerns among governments about the rapid evolution of artificial intelligence, which threatens to overwhelm regulatory frameworks governing military uses of artificial intelligence. Such a trajectory increases the likelihood of accidents, miscalculations, or unexpected escalations.
Dutch Minister of Defense Ruben Brekelmans articulated the dilemma facing the government, metaphorically comparing it to a “Prisoner’s Dilemma.” He observed that states are caught between the need to impose responsible constraints and avoiding self-imposed limitations on adversaries.
“Russia and China are moving forward quite quickly,” Brekelmans told Reuters. “This is forcing us to accelerate our AI development efforts. But this acceleration also increases the urgency of promoting the responsible use of AI. These two imperatives are inherently interconnected.”
Just 35 of the 85 countries gathered at the Responsible AI in the Military (REAIM) Summit in La Coruña on Thursday supported commitment to the 20 Guiding Principles on AI.
These principles include affirming human responsibility for the operation of AI-enabled weapons, promoting transparent command and control structures, and facilitating the exchange of information on national oversight mechanisms consistent with national security interests.
Additionally, the document emphasized the need for risk assessments, rigorous testing, and comprehensive training and education for personnel working with military AI systems.
At previous military AI summits held in The Hague and Seoul in 2023 and 2024, about 60 countries, including the United States but excluding China, endorsed a preliminary “blueprint for action” that lacked binding legal commitments.
While the current text maintains a non-binding position, some participants expressed discomfort in favor of a clearer policy, as articulated by Yasmin Afina, a researcher at the United Nations Institute for Disarmament Research and an advisor involved in the discussions.
Major signatories to the recent agreement include Canada, Germany, France, the United Kingdom, the Netherlands, South Korea and Ukraine.
Source link: Yahoo.com.
