United Nations-backed talks in Geneva concluded on Saturday, November 21, 2026, with an agreement to extend discussions on the regulation of AI-powered weapons for two more years. Scores of participating countries pledged to elaborate a possible accord on lethal autonomous weapons systems by 2028, but they did not commit to legally binding rules requested by U.N. officials and humanitarian organizations.
The week-long conference operated under the Convention on Certain Conventional Weapons, a treaty ratified by 128 countries that requires decisions to be made by consensus. This structure allows any single nation to block an agreement. U.N. Secretary-General António Guterres, who leaves office at the end of 2026, had previously set a 2026 deadline to launch negotiations for an agreement.
During the session, seven countries—Austria, Belgium, Brazil, Ireland, Mexico, New Zealand, and Panama—proposed starting negotiations on a rulebook to limit machines' ability to kill without human accountability. This effort fell short, with advocacy groups stating that military powers like the United States and Russia blocked stronger language. Advocacy groups including Stop Killer Robots and Lex International criticized the outcome as a watered-down mandate.
Around 100 states supported legally binding rules, but the lack of a firm commitment means there is currently no international law specifically prohibiting machines from identifying and engaging targets without human oversight. Proponents of regulation argue this delay allows technology to outpace diplomacy.
The two-year extension allows for the continued development of these technologies. What happens next is a series of continued discussions with a target date of 2028 for nations to elaborate on the details of a possible international agreement.
