Netherlands asks UN members to ban AI-controlled nukes

Dutch Prime Minister Dick Schoof on Wednesday presented guidelines for artificial intelligence use in the military to the United Nations General Assembly. He urged UN member states to sign a legally binding agreement to keep the use of nuclear weapons under human control. Such weapons require moral, legal, and strategic considerations beyond the scope of AI, the guidelines said. The guidelines were compiled by a global commission in conjunction with the Dutch government and The Hague Center for Strategic Studies.

UN Secretary-General Antonio Guterres agreed with Schoof, calling for a global ban on AI-controlled lethal weapons systems during the session. He asked UN members to consider signing such a ban by next year, according to UN News.

What other recommendations did Schoof present? He advocated for the continuous training of humans who use AI for military purposes, so that they could remain in control of the technology and understand how it operates. Military AI systems should also be developed so that a human always has final responsibility for critical decisions, the guidelines said. Schoof also recommended that UN states develop a centralized network of experts who would share AI development information. They would coordinate policies and regulations with the UN, according to the guidelines.

How did the United States respond to the recommendations? The U.S. government totally rejected all attempts to institute global governance of AI, White House Office of Science and Technology Policy Director Michael Kratsios told attendees of the UN session. Overregulation and a fixation on social equity would stifle technological development, he said. The United States would keep the use of AI in line with Americans’ civil liberties and would focus on exporting technology to its allies, he said.

Dig deeper: Read Carolina Lumetta’s report on President Donald Trump’s speech to the United Nations.