At a United Nations meeting in Geneva in November 2017, attendees of the Convention on Certain Conventional Weapons—a group focused on lethal autonomous systems—watched a short film called Slaughterbots. The video depicted an AI-powered drone capable of killing targets unassisted. "People don't," the fictional CEO says in the film. "They get emotional, disobey orders, aim high. Let's watch the weapons make the decisions." Branka Marijan, a senior researcher at Project Ploughshares who was in the room, recalled the mood shifting to immediate apprehension. The most unsettling realization wasn't the premise itself—it was that the Pentagon was already developing similar technology. That UN gathering marked the first meeting held after the launch of Project Maven, a US Department of Defense initiative using AI to analyze drone surveillance footage. By late 2017, Google had signed on as a major tech partner for the project. Marijan emphasized that the systems being discussed were far from futuristic—they were existing platforms with degrees of autonomy already built in, capable of selecting and engaging targets based on sensor data and input. The convention, which convenes twice yearly in Geneva, had historically dealt in hypotheticals about a world of killer robots. But in 2017, Marijan realized the distant, imagined future had become far closer and more tangible. The world had already witnessed drone warfare in various forms, and the addition of AI capabilities was transforming what those systems could do without human intervention.