U.S. and China Skip New AI Weapons Pledge in Spain as Only 35 Nations Sign

February 6, 2026
U.S. and China Skip New AI Weapons Pledge in Spain as Only 35 Nations Sign

A CORUNA, Spain, Feb 6, 2026, 18:21 (CET)

  • U.S. and China declined to join a new declaration on AI use in warfare at a Spain summit
  • Just 35 of 85 countries attending signed a voluntary set of 20 principles
  • Delegates cited strained U.S.-Europe ties and uncertainty over future transatlantic relations

The United States and China opted out of a joint declaration on the military use of artificial intelligence at a summit in Spain on Thursday, leaving about a third of participating countries signed up to a set of voluntary principles. Reuters

The pledge matters now because governments are trying to bolt guardrails onto a fast-moving technology that is already being tested and fielded in defence settings. Supporters of the declaration warn that if rules lag, AI in warfare could raise the odds of accidents, misreads and unintended escalation.

The politics are tangled, too. Several attendees said recent strains between Washington and European allies — and doubts about how those ties will look in coming months and years — made some states cautious about signing anything at all.

The declaration, agreed at the Responsible AI in the Military Domain (REAIM) meeting in A Coruna, sets out 20 principles aimed at steering how militaries deploy AI in conflict. It stresses that humans remain responsible for the use of AI-enabled weapons, calls for clear lines of command and control, and urges countries to share oversight arrangements when it does not clash with national security. Dawn

It also points to risk assessments, robust testing and training for personnel operating military AI capabilities, according to the text described by delegates.

Dutch Defence Minister Ruben Brekelmans said governments face a “prisoner’s dilemma” — wanting restraints, but fearing rivals will not play along. “Russia and China are moving very fast,” he said, adding that speed in development and responsible use “go hand-in-hand.”

Earlier REAIM meetings in The Hague in 2023 and Seoul in 2024 drew backing from around 60 countries for a looser “blueprint for action” that carried no legal weight. China stayed out of those documents, but the United States signed then. Indiatimes

This year’s declaration is also non-binding — not legally enforceable — yet some governments were still uneasy about putting their name to a more concrete set of policies, said Yasmin Afina, a researcher at the U.N. Institute for Disarmament Research who advised the process.

Countries that signed on Thursday included Canada, Germany, France, Britain, the Netherlands, South Korea and Ukraine.

The low take-up, and the absence of Washington and Beijing, left the summit with a familiar problem: broad agreement that AI brings new military risks, and much weaker agreement on who should limit what, and when.

But the bigger uncertainty is what, if anything, the pledge changes. With no enforcement, and with only 35 signatories out of 85 attendees, the declaration may struggle to shape behaviour if major military powers keep their options open.

The AI Arsenal That Could Stop World War III | Palmer Luckey | TED

Technology News

  • AI deepfake romance scam drains woman's life savings and home
    February 6, 2026, 12:34 PM EST. A woman identified as Abigail believed she was in a romance with a famous actor after messages and an AI-generated video that looked and sounded real. The scam began on Facebook, moved to WhatsApp, and used a convincing deepfake video to seal trust. Her daughter, Vivian Ruvalcaba, says the case unfolded quietly; grooming and isolation kept the victim from telling family. By the time relatives intervened, more than $81,000 was stolen and the paid-off home she planned to retire in was at risk. The episode illustrates how AI tools can fuel romance scams by mimicking a real person, including impersonations of actors like Steve Burton. Authorities were alerted after the manipulation escalated to financial loss.