A Senate committee has adopted an amendment from Arizona Sen. Mark Kelly, aimed at requiring "ultimate human responsibility" in the use of force within AI-powered kill chains. This measure codifies the 2023 Defense Department policy, Directive 3000.09, which mandates human oversight for autonomous weapons systems. The amendment directly challenges President Donald Trump's June 5 memorandum, which sought to "eliminate unnecessary barriers to rapid deployment" of AI and update Directive 3000.09
within 90 days, potentially removing human involvement. The U.S. has seen escalated AI use in conflicts, including the war with Iran and the Russia-Ukraine war. For instance, the Maven Smart System, engineered by Palantir Technologies, was implicated in a missile attack in Iran that killed 175 people, mostly young girls, reportedly due to outdated targeting data. Experts like Zaza Tsotniashvili and Paul Lushenko stress the legal and ethical necessity of human oversight, arguing that DoD directives alone are insufficient without congressional codification and robust accountability mechanisms to prevent the erosion of meaningful human control.
No comments:
Post a Comment