Musk-Backed 'Slaughterbots' Video Will Warn the UN About Killer Microdrones (space.com)

An anonymous reader quotes Space.com: A graphic new video posits a very scary future in which swarms of killer microdrones are dispatched to kill political activists and U.S. lawmakers. Armed with explosive charges, the palm-sized quadcopters use real-time data mining and artificial intelligence to find and kill their targets. The makers of the seven-minute film titled Slaughterbots are hoping the startling dramatization will draw attention to what they view as a looming crisis -- the development of lethal, autonomous weapons, that select and fire on human targets without human guidance.

The Future of Life Institute, a nonprofit organization dedicated to mitigating existential risks posed by advanced technologies, including artificial intelligence, commissioned the film. Founded by a group of scientists and business leaders, the institute is backed by AI-skeptics Elon Musk and Stephen Hawking, among others. The institute is also behind the Campaign to Stop Killer Robots, a coalition of non-governmental organizations which have banded together to call for a preemptive ban on lethal autonomous weapons... The film will be screened this week at the United Nations in Geneva during a meeting of the Convention on Certain Conventional Weapons... The Campaign to Stop Killer Robots is hosting a series of meetings at this year's event to propose a worldwide ban on lethal autonomous weapons, which could potentially be developed as flying drones, self-driving tanks, or automated sentry guns.
"This short film is more than just speculation," says Stuart Russell, a U.C. Berkeley considered an expert in artificial intelligence.

"It shows the results of integrating and miniaturizing technologies we already have."

Comments

Re:The problem is they're too cheap

[mentil]

Forget artillery strikes. What's really gonna stir shit up is when drone attacks allow for anonymous murder. What society could cope with that? It's gonna be like handing a Death Note to every citizen on Earth, only you won't need to know their name. I COULD see a ban on murder-drones actually working though, as after the first time a crime family gets whacked, the black market won't touch them; people will go after individual sellers as well. Wearing a mask in public might become the norm... until voice-recognition is used instead.

Re:A firecracker, not a bomb

[Kjella]

A more realistic weapon for a palm-sized drone would be a wasp-style injection, just ram the target and have a spring released needle to punch through the clothing and skin or shoot from close range like a taser gun. There are plenty toxins you could deliver that would be fatal with even a very tiny dose. But the topic wasn't really if a drone attack is practical, but whether an autonomous drone attack is more practical. I don't really see it, if there's only one target then human RC will do fine. If there's many targets, gathered in a relatively small space like some form of meeting or conference, why wouldn't you just hit that with one big bang? I mean the assumption here is that you're willing to commit mass murder, are you going to care if there's a little collateral?

I'd think the only reason you'd care is because you're trying to be the good guys, like IS is using human shields or they're in a camp with women and children or whatever and you want to make precision kills without harming the rest. In any case there's probably good reasons to build all this technology for non-lethal purposes, swapping out some non-critical function with some kind of trigger/detonator is always going to be easy. It's like trying to build an alarm clock that can't be rigged to blow up a bomb when the alarm goes off. What are you going to do, ban alarm clocks?