The last two decades saw the experiments and the fielding of various drones, large and small, armed and unarmed. There has been some ethical chitchat about the use of these, but without much echoing in the media. The drones that has been used lately to kill people (a.k.a. “neutralize high value targets”) were controlled by human operators and the orders to release ordnance were also given by humans. Or so was the situation until recently.
Because of the nature of targeted killings few if any real information is available, yet it is clear that at least the major nation states have ongoing programs to develop and operate semi-autonomous machines to deal with scenarios where the goal is to neutralize a selected individual. If you take this out of the context, it bears some ( 🙂 ) resembleance to the oldschool job of a hitman, making it difficult to differentiate between the ethos of the governments and say, transnational criminal organizations.
The inhumanity of the formal wording of contract killing is only concealed by use of the word ‘insurgent’ – and nothing else.
High-Value Targeting is defined as focused
operations against specific individuals or networks
whose removal or marginalization should
disproportionately degrade an insurgent group’s
effectiveness. The criteria for designating high-
value targets will vary according to factors such as
the insurgent group’s capabilities, structure, and
leadership dynamics and the government’s desired
outcome. (excerpt from a CIA manual)
But if that was not bad enough, here is the new way to kill your select enemies: a swarm of air-launched drones. In a controlled experiment in 2016, a pair of US Navy McDonnell Douglas F/A-18F two-seat Hornet fighter-bombers released 104 Perdix drones (that is 26 per container), which then autonomously moved to waypoints set by human operators. The machines involved has been around for a while: the F/A-18 Hornet‘s maiden flight was on 9. June 1974 (in the form of the YF-17 / P530 “Cobra”) and the Perdix intelligent micro-drone was designed in 2011 at the Aeronautics and Astronautics Department of the Massachusetts Institute of Technology (MIT).
The first military model of Perdix, has been modified in 2013 for the US DoD. The specifications are not known, but it is obvious that these micro-drones are able to communicate with each other, are able to share the workload and are able to perform after being launched from a container carried on a wing pylon of a jet fighter.
According to the information released, the frames of the Perdix drones are being built at least partly with 3D printers and they are able to function for more than 20 minutes. According to scientific media reporting, “the wings are made of carbon fiber and the fuselage is made of a kevlar composite. The drone is powered by a lithium polymer battery pack powering a rear-facing push propeller”. A quick calculation leads us to the conclusion that the range of the micro-drones if used in a non-reuseable way might be 30-50 kilometres depending on the launch altitude and its abilities to descend without full propulsion.
What is not clear for the casual reader is that these are actually built around simple hand-held mobile devices, which are the ‘brains’, ‘souls’ and ‘senses’ of the micro-drones. And the capabilities of a single smartphone are more than enough to power an autonomous machine.
The potential military use of the Perdix swarm seems clear: reconnaissance of the enemy, jamming enemy communications, providing a forward signal repeater network for long-range special operations by retransmitting signals, delivering small explosive charges (this has been done even by the ISIS, so no news here either) and act as decoys spoofing enemy radars (by something like a diameter-enlarging technique or dispensing chaffs and/or flares perhaps).
Based on the redacted news that are from two years ago, we don’t know for sure how autonomous the swarm or the individual micro-drone actually is. It is possible that it cannot function without receiving objectives from a human controller beforehand, but it is equally possible that it will pursuit its pre-set objectives once launched, no matter what.
The fact that all the enthusiasm about the Perdix drone swarm is centered around its demonstrated ability that it had been told what to do, but not how to do it makes us think that this phenomenon is actually a neural network of sorts or in other words: a working autonomus AI based on flying smartphones. Which is scary to say the least.
For a quick look at the experiment, here’s a video: