How far are we from AI robots killing people based on algos?
few years or a decade? It's probably possible to build them today but I talk about millions of them?
Looking at the drone war in Ukraine there is an intense battle to disrupt control signals to drones. If the drone can drive it self and take action without input, the effectiveness of signal jamming becomes much more reduced.
Patriot missile system can mark, track, and kill a pilot violating controlled airspace entirely autonomously: https://en.wikipedia.org/wiki/MIM-104_Patriot
Just shy of "make it humanoid with digital arms and legs controlled by servos", lethal autonomy has been operationalized in armies worldwide for decades now. It's scary shit: https://youtu.be/bg9uoI8RQKc