Killer Robots

Fathom

Drones occupy battlefields today. Airborne. Strike targets. Automated warfare exists today. This technology said to be an expanding "continuum." AI weapons development replacing direct human involvement. Should limits be placed on such “lethal autonomous weapons?” Computer scientist Stewart Russell at University of California at Berkeley thinks so. But is it too late? Ask I-Robot? Terminator? Sci Fi leads the way? Ref: http://www.popsci.com/big-idea-killer-robots-are-coming

Capt'n Midnight

From of all things a Chuck Norris film Code of Silence. It featured a Prowler Robot with machine guns , laser range finders and infra red night vision back in 1985. Product placement of a real murder machine.

Promo here https://www.youtube.com/watch?v=GfT96j-7Zjc
Movie Kill-count (so NSFW) the first robot kill was with Chuck at the remote control, can't remember if the second ones were, but after that it was autonomous, it was kinda sickening watching it the first time because it wasn't a human doing the killing.
https://www.youtube.com/watch?v=E9dqfCvKC1A

These Prowlers existed before the first Terminator movie. But back then no one was buying them. We have entered the fourth decade of functional autonomous killer robots.




http://explainxkcd.com/652/

We live in a world where there are actual fleets of robot assassins patrolling the skies. At some point there, we left the present and entered the future.

ScumLord

Fathom wrote: »

Should limits be placed on such “lethal autonomous weapons?”

By who? America wouldn't give up it's advantage and Russia wouldn't want to be told they can't match the US on military might.

At this stage a super intelligent AI taking over is probably the best case scenario for humanity. We keep putting our animal instincts and human emotions onto AI and we don't seem to be able to imagine past them being just extra human. It would have none of those hangups and there's no reason to think it will want to kill everything by default.

4068ac1elhodqr

ScumLord wrote: »

At this stage a super intelligent AI taking over is probably the best case scenario for humanity. .

i) Artificial Narrow Intelligence, has been around for a fair while
ii) Assisted Artificial General Intelligence is the current, and 'aspired to' standard.

iii) Artificial 'Super-intelligence', (most likely through some sort of H+ organic neural interfacing) is still some time away, perhaps 2050++.
So of little use to us between now and then.

One concern is the new breed of robotics that may be able to self-reproduce and even sustain themselves upon organic matter.

ScumLord

I'm not really too worried about robots until there's something independant to run them without human help.

So much sci-fi gives the impression that we'll accidentally make robots that will just spontaneously become smart and hate us as a consequence. Not only do they become sentient but they become human sentient, as if that's the only state of consciousness that matters, or can exist.

I don't really see an automated system just deciding to run amok. When our tech fails it doesn't fail in a coordinated way, it just stops working. At worst the AI will just make the same mistakes as people, like incorrectly identifying targets, it would probably just do it less than people. It won't get afraid that a child might have a bomb, it will just register a non combatant and leave it at that.

Bottom line is machines are far from self sufficient. They need people in so many ways that any attempt to turn against the will of people will fail pretty rapidly.

Fathom

Sending humans to Mars? Will flight inventory include weaponry? Armed robots? Spare humans. Save oxygen payload. Send only robots. (Insert Star Trek meme. To explore...)

Jurgen Klopp

China and several others are already calling for a international treaty on the Geneva convention level to outlaw any that can operate without human control.

So like drones being flown from what appears to be an Xbox remote from base a possible future terminator like soldier woukd need to also be controlled by a "pilot" of sorts.

Apparently so there can be no washing of hands if one killed a civilian. I suppose in a nation would claim software error due to battle damage or some BS. Also that it means humans always remain in control.

Riesen_Meal

Jurgen Klopp wrote: »

China and several others are already calling for a international treaty on the Geneva convention level to outlaw any that can operate without human control.

So like drones being flown from what appears to be an Xbox remote from base a possible future terminator like soldier woukd need to also be controlled by a "pilot" of sorts.

Apparently so there can be no washing of hands if one killed a civilian. I suppose in a nation would claim software error due to battle damage or some BS. Also that it means humans always remain in control.

Jurgen, you have a matter of a Derby to be planning, get off Boards... :P

Fathom

Jurgen Klopp wrote: »

China and several others are already calling for a international treaty on the Geneva convention level to outlaw any that can operate without human control.

Film I-Robot. Can robots murder? Current legal definitions?

Fathom

ScumLord wrote: »

By who? America wouldn't give up it's advantage and Russia wouldn't want to be told they can't match the US on military might.

Another Cold War (robot) arms race?

Capt'n Midnight

Fathom wrote: »

Film I-Robot. Can robots murder? Current legal definitions?

At the very least I consider it manslaughter.

Once the AI goes live it's the same as arming a landmine.

Fathom

Robots injure or kill. Workplace. "Accident" or what? Proving criminal law "intent" vs civil law damages? Can AI be held responsible? Human owner? Programmer? All the above? Interesting legal questions. Ref: http://www.nytimes.com/interactive/2014/06/16/upshot/100000002939565.mobile.html?abt=0002&abg=0&_r=1

Capt'n Midnight

Fathom wrote: »

Robots injure or kill. Workplace. "Accident" or what? Proving criminal law "intent" vs civil law damages? Can AI be held responsible? Human owner? Programmer? All the above? Interesting legal questions.

Ref: http://www.nytimes.com/interactive/2014/06/16/upshot/100000002939565.mobile.html?abt=0002&abg=0&_r=1

I don't thin AI was involved in any of those. Sounds like all those incidents could have been preventable with adequate safety switches or procedures or staying out of the well marked danger area.

Fathom

Capt'n Midnight wrote: »

I don't thin AI was involved in any of those. Sounds like all those incidents could have been preventable with adequate safety switches or procedures or staying out of the well marked danger area.

Agree. Sci Fi Skynet may be in the distant future. If at all.

ban resistant recalcitrant debutant

A self driving car is a robot and it could kill someone.

edit.
I didn't realise the thread was so old.

About the car, if a self driving car today did kill someone in an accident what information would be available about the crash?

Are there black boxes mandated by law that self driving cars must have?

There should be really, and the black box should contain information from the sensors, and information about what decisions were made by the car and why.

Fathom

Experts call for ban on killer robots.

EPAndlee

Has I Robot,Terminator etc thought us nothing

Fathom

EPAndlee wrote: »

Has I Robot,Terminator etc thought us nothing

Jules Verne's 20,000 Leagues Under the Sea theme.

Black Swan

Fathom wrote: »

Experts call for ban on killer robots.

The Military Industrial Complex will win and receive killer robots given their vast monies and congressional lobbies. Upon reflection, it's already a fait accompli today with flying drones attacking military targets in Yemen, etc.

Fathom

Black Swan wrote: »

Upon reflection, it's already a fait accompli today with flying drones attacking military targets in Yemen, etc.

Drones controlled thousands of miles away from target by real military game players.

Fathom

Fathom wrote: »

Drones controlled thousands of miles away from target by real military game players.

Boon for gaming industry?

Fathom

Free classes on Autonomous Mobile Robots provided by edX.

josip

As soon as this guy gets his whatevers on an AK47 we're done.

https://newatlas.com/atlas-robot-backflip/52234/

4068ac1elhodqr

Agree this (darpa's) new Atlas2 chap is a bit of a 'leap' ahead of previous models that stumbled over it's wires.
https://www.youtube.com/watch?v=fRj34o4hN4I

Musk reckons this (ai-bots) are even more of a risk than Rocketman himself.

Fathom

Free edX Artificial Intelligence courses.

Capt'n Midnight

Looks like the start of a trend, a way for corporations to make money.

https://www.bizjournals.com/sanfrancisco/news/2017/12/08/security-robot-homeless-spca-mission-san-francisco.html

, the security robot, which they've dubbed K9, was a way to try dealing with the growing number of needles, car break-ins and crime that seemed to emanate from nearby tent encampments of homeless people along the sidewalks.

...
The robot costs about $6 per hour to rent,

Fathom

Delivery robots. Security robots. Regulation soon to follow.

Fathom

Fathom wrote: »

Delivery robots. Security robots. Regulation soon to follow.

San Francisco implements tight regulation on delivery robots.

Fathom

Fathom

Matt Simon in December Wired: 2017 Was the Year the Robots Really, Truly Arrived.

Black Swan

The companion robots of science fiction went on sale in US December 2017. What's next? Will Smith's i-robots? Or a bit later, a Alicia Vikander Ex Machina?

Fathom

Was Ex Machina a "killer robot?"

Black Swan

Fathom wrote: »

Was Ex Machina a "killer robot?"

See the film.

Capt'n Midnight

Black Swan wrote: »

See the film.

Or better still , watch Klaus Kinski do his thing in the low budget Android (1983)

I tried to watch Ex Machina and after a while just had to stick it on fast forward









The problem with robots is people.

srsly78

Black Mirror Season 4 Episode 5 - Metalhead

Most realistic depiction of killer robots to date?

Fathom

srsly78 wrote: »

Black Mirror Season 4 Episode 5 - Metalhead Most realistic depiction of killer robots to date?

Spoiler alert: Link to above.

ScumLord

ScumLord wrote: »

I'm not really too worried about robots until there's something independant to run them without human help.

So much sci-fi gives the impression that we'll accidentally make robots that will just spontaneously become smart and hate us as a consequence. Not only do they become sentient but they become human sentient, as if that's the only state of consciousness that matters, or can exist.

I don't really see an automated system just deciding to run amok. When our tech fails it doesn't fail in a coordinated way, it just stops working. At worst the AI will just make the same mistakes as people, like incorrectly identifying targets, it would probably just do it less than people. It won't get afraid that a child might have a bomb, it will just register a non combatant and leave it at that.

Bottom line is machines are far from self sufficient. They need people in so many ways that any attempt to turn against the will of people will fail pretty rapidly.

When technology starts malfunctioning it doesn't just stop working though. A small error, either on its part or by human interference can cause it to continue doing what its programmed to do, til the end of its program with sometimes unexpected outcomes.

Face recognition software has been tricked by printouts of faces. What if a drone is given an order to kill a dissident who is also running for election in a country that has posters with his face up all over the place and the drone strafes whole apartment buildings?

dr.fuzzenstein

denartha wrote: »

When technology starts malfunctioning it doesn't just stop working though. A small error, either on its part or by human interference can cause it to continue doing what its programmed to do, til the end of its program with sometimes unexpected outcomes.

Face recognition software has been tricked by printouts of faces. What if a drone is given an order to kill a dissident who is also running for election in a country that has posters with his face up all over the place and the drone strafes whole apartment buildings?

It should be precise enough to just destroy the posters.

Black Swan

Capt'n Midnight wrote: »

Or better still , watch Klaus Kinski do his thing in the low budget Android (1983)

Or Ash in Alien (1979); or Bishop in Aliens (1986) and Alien 3 (1992)?

4068ac1elhodqr

/ Lifesaving robots.

(Whilst not true AI), a small drone has saved two swimmers down in Aussieland.
From launch to accurate device drop in about 70secs.

"Never before has a drone, fitted with a flotation device been used to rescue swimmers like this."

Capt'n Midnight

Archer Eager Viper wrote: »

From launch to accurate device drop in about 70secs[

200 years ago they'd have used a Manby mortar or a rocket ( or later on a cannon ) to deliver the line.

Black Swan

Hospital delivery robot rolls out in Japan.

Fathom

Black Swan

Above definitely fulfills the Killer Robots subject.

Fathom

As does this flying one.

Fathom

Robot army. United States Army's Future Combat Systems Project. Use of artificially intelligent systems in modern warfare. Behind original deployment schedule. Alex Leveringhaus in What's So Bad About Killer Robots? (Journal of Applied Philosophy, 6 March 2016), questions if engaging a target without supervision by a human operator should be allowed.

Black Swan

Killer robots: Philosophical ethics understood only by a few vs the Military Industrial Complex profits under Trump. I can just imagine who would win.

Fathom

"Profits not people" maxim? Specifically killer robots not people.

Professor Moriarty

So, could/should a robot be put on trial for war crimes?

Fathom

Professor Moriarty wrote: »

So, could/should a robot be put on trial for war crimes?

Did movie i-Robot consider robot rights to trial?

Professor Moriarty

Fathom wrote: »

Did movie i-Robot consider robot rights to trial?

Dunno. Just wondering in general in the context of free will.