Stop Autonomous Weapons
4 HN recommenders · recommended in 16 months, 2018–2026
7:48 · 3.9M views · Nov 12, 2017
Slaughterbots is a fictional short film, backed by Stuart Russell, depicting swarms of cheap autonomous drones that use facial recognition to assassinate targets. It was made to argue for banning autonomous weapons.
HN invokes it as the standard reference whenever drone swarms, autonomous weapons or military AI come up in the news.
20 comments from Hacker News
Can drones be stopped? They have the potential to wage campaigns of automated political violence. I see companies trying to protect against them, like Dedrone [1] -- how effective are these? I imagine most here have seen it, but if not, you must see the slaughterbots video. [2] [1] [LINK] [2] this video
The substantive part of this conversation comes about 35 minutes in, when Allen describes the DoD's five AI ethics principles: that any deployed AI systems must be "Responsible, Equitable, Traceable, Reliable, and Governable." [1]. I think "Responsible" raises some hard questions. Today, any weapons use by a robot might require a responsible human operator to first (remotely) pull the trigger. But what if an adversary allows their robots to fire without waiting for human permission? The first side to do so would gain a critical advantage, and we cannot have an autonomy gap. Some have raised conerns that in the extreme case, this could lead to a world of machines anonymously killing without attribution. If you haven't, see Stuart Russell's compelling short film "Slaughterbots" [2]. [1] [LINK] [2] this video
The tech required for search and rescue is the same tech required for search and destroy. We seem to be running headfirst into the autonomous killer bot future without giving so much as a second look at the implications. Slaughterbots [0] is a pretty good look at what can be achieved by refining and integrating technologies which all exist today. [0] this video
This brings to mind a great video which I've posted here a few times: this video What happens when technology progresses so much that anyone can make a swarm of killer drones? Will politicians no longer be able to go out in public? What about a distant future, when enriching uranium can be done in your garage for cheap...
I don't think any of tensions, classical conflicts, asymmetric conflicts, the amount of people with nothing left to "loose", or all of these together, are at particularly unusual levels compared to the previous 6000 years, though they're higher than they were 20 years ago. But, for most of that time, warfare was geographically localized; you could avoid directly experiencing warfare by not being a soldier and living inside a country that wasn't being actively invaded, or on the national border between two countries that weren't at war with each other or being actively invaded. Sometimes that was easier said than done, but most people managed it most of the time. In the Drone Age, though, you can remotely pilot a quadcopter 4300 km away to blow up airplanes on the airstrip: [LINK] — or to blow up your political opposition, if you can guess where they are or will be. The US has been doing this for 20 years in Afghanistan and Iraq: [LINK] (use Readability mode to bypass "this is not a paywall") but a General Atomics MQ-9 Reaper costs on the order of US$100 million, so there are less than 400 of them [LINK] while Ukraine's most popular drones cost on the order of US$300, can be 3-D printed in a basement, already cause 70% of casualties on the battlefield, and are produced in volumes approaching 10 million per year [LINK] . Many experts believe drones have made tanks obsolete on the battlefield [LINK] . Even without autonomous weapons, we're rapidly moving toward the future of bor