The UN Will Finally Decide In 2017 If Armies Can Use ‘Killer Robots’

[annabenedetti]

The UN is set to tackle the issue of killer robots amid concerns from Silicon Valley leaders and human rights organisations over the threat they pose to humanity.

Delegates representing 123 nations at the Convention on Conventional Weapons have agreed to form a group next year to address their use.
Human Rights Watch said the meeting marked a step “towards a ban” on weapons that can strike without human intervention.

. . . .

Dozens of artificial intelligence experts signed a letter last year, calling for an outright ban on the development of AI weapons.
The letter warned that killer robots would lower the threshold for going into battle and could help subdue populations and aid acts of genocide.

“The governments meeting in Geneva took an important step toward stemming the development of killer robots, but there is no time to lose” said Steve Goose, arms director of Human Rights Watch, a co-founder of the Campaign to Stop Killer Robots.

“Once these weapons exist, there will be no stopping them. The time to act on a pre-emptive ban is now.”

In the letter, the experts warned: “Autonomous weapons are ideal for tasks such as assassinations, destabilizing nations, subduing populations and selectively killing a particular ethnic group. We therefore believe that a military AI arms race would not be beneficial for humanity.”

 

[annabenedetti]

I was reminded of this:

This robot is the height of an average man. I don't want to envision a vast army of these that are bigger, stronger, and more advanced. Not to mention the capabilities of drones equipped with biological weapons.

 

[ok doser]

be nice if they managed to get it right with driverless cars first, doncha think?

 

[annabenedetti]

They've been using robots in combat and law enforcement for what, a decade or more? In some ways they save lives, and in some ways they take lives. Their use raises ethical questions that are going to have to be addressed, and sooner rather than later because the future is already here.


And here's an excerpt from an interesting discussion in the aftermath of the police robot killing of the Dallas Police shooter:

How does this link to the "killer robots" debate?

While there are links to the discussion on the future of killer robots, known as LAWS (lethal autonomous weapons systems), there are also three important differences to keep in mind. First, this was not an autonomous robot; both the robot and the explosive were remotely operated. Second, this was not a designed weapons system. There are many ground robots under development that are armed with weapons like machine guns and missile launchers, from the prototypes of the US MAARS to the Chinese Sharp Claw. But this was the case of using an older robot designed for something else. And, third, the concerns driving the "killer robots" debate center on whether we can prevent the unleashing of self-operating robots on the battlefield, for which militaries around the world are beginning to establish plans and doctrines. The debate is driven by a concern that such a future might cause both greater civilian harm and a lack of accountability.​

 

[glassjester]

The UN is set to tackle the issue of killer robots amid concerns from Silicon Valley leaders and human rights organisations over the threat they pose to humanity.

I'm 100% for a ban.

But how long could the ban last?

 

[ok doser]

The UN is set to tackle the issue of killer robots amid concerns from Silicon Valley leaders and human rights organisations over the threat they pose ...

so we're gonna use killer robots to destroy silicon valley leaders and human rights organizations?


sweet! :thumb:

 

[rexlunae]

They've been using robots in combat and law enforcement for what, a decade or more? In some ways they save lives, and in some ways they take lives. Their use raises ethical questions that are going to have to be addressed, and sooner rather than later because the future is already here.


And here's an excerpt from an interesting discussion in the aftermath of the police robot killing of the Dallas Police shooter:

How does this link to the "killer robots" debate?

While there are links to the discussion on the future of killer robots, known as LAWS (lethal autonomous weapons systems), there are also three important differences to keep in mind. First, this was not an autonomous robot; both the robot and the explosive were remotely operated. Second, this was not a designed weapons system. There are many ground robots under development that are armed with weapons like machine guns and missile launchers, from the prototypes of the US MAARS to the Chinese Sharp Claw. But this was the case of using an older robot designed for something else. And, third, the concerns driving the "killer robots" debate center on whether we can prevent the unleashing of self-operating robots on the battlefield, for which militaries around the world are beginning to establish plans and doctrines. The debate is driven by a concern that such a future might cause both greater civilian harm and a lack of accountability.​

I'm not too worried about that case. It's a robot, but not an automaton. If we allow independent robotic minds to just wage war on their own initiative, there's a huge potential for that to get away from us, even assuming that it's judged to be a reasonable tactic to allow in warfare. Human minds are dangerous enough.

 

[Quincy]

The problem I have with this, as a person who's played a lot of computer deathmatches featuring bot a.i.s , is that programmers usually create an a.i. that does one of two things. Either they are so lethal and precise in execution that they kill everyone including their own allies or they are so dumb that they just chest-bump the enemy for 20 minutes or try to snipe with a shotgun. Seems like it's incredibly difficult to program an a.i. that is somewhere in the middle.

I'd personally like to see a war featuring the derpy robots, though.

 

[The Berean]

I think people have been watching way too many sci-fi movies. We are a long, long way from having Terminator style robot armies, if ever. In 2015 DARPA hosted a robotics competition. The competition was a series of tasks a robot must complete. These were a series of tasks that any able bodied person could complete in 15-20 minutes. Most of the robots failed miserably.

http://www.roboticstrends.com/article/14_epic_fails_from_the_darpa_robotics_challenge_finals



Sent from my SM-G920V using TheologyOnline mobile app

 

[rexlunae]

The problem I have with this, as a person who's played a lot of computer deathmatches featuring bot a.i.s , is that programmers usually create an a.i. that does one of two things. Either they are so lethal and precise in execution that they kill everyone including their own allies or they are so dumb that they just chest-bump the enemy for 20 minutes or try to snipe with a shotgun.

The bots you find in a computer game are barely or not-at-all real AIs. They're usually programmed with pretty simple behavior. A real AI learns and adapts like a real intelligence.

Seems like it's incredibly difficult to program an a.i. that is somewhere in the middle.

You don't really program an AI. Of course, you have to write some code, but mostly you train an AI.

 

[The Berean]

I was reminded of this:

This robot is the height of an average man. I don't want to envision a vast army of these that are bigger, stronger, and more advanced. Not to mention the capabilities of drones equipped with biological weapons.

I wouldn't worry too much about this robot, Anna. The company that created this robot, Boston Dynamics, had been in business for 24+ years. To this day Boston Dynamics has yet to commercialize a single robot. Google purchased them a while back then this past May announced they were selling the company. A major reason for the sale is that Google felt Boston Dynamics was not really close to having a viable commercial product in the near term future. Boston Dynamics is great at creating fancy prototypes and making cool videos but it's quite obvious their prototypes are of little practical use. They have lived off DARPA/DOD money for years and it's quite telling that the US Military has never made any operational use of their robots.

Sent from my SM-G920V using TheologyOnline mobile app

 

[annabenedetti]

I'm not too worried about that case. It's a robot, but not an automaton. If we allow independent robotic minds to just wage war on their own initiative, there's a huge potential for that to get away from us, even assuming that it's judged to be a reasonable tactic to allow in warfare. Human minds are dangerous enough.

Yes, he made the distinction between that robot (human-operated, already existing and retro-fitted for a single job) and the future of military robots. But the potential ethical dilemmas are already being considered, and have been considered, for example, in our use of drones. But again, for now, the drones are operated by humans. But everything is moving forward so fast, and AI particularly in deep learning and neural network seems to be moving at warp speed. I was reading recently about companies snapping up academic AI experts and graduating AI engineers as fast as they could get them, offering graduates seven-figure salaries. Unfortunately I didn't save it and can't find it at the moment to link it here, but here's a different look at how fast things are advancing at companies such as Google.

I don't know a lot about the deep learning and neural networks of AI, but I pay attention to what I read and try to read from good sources, because it's my belief that AI may move faster than human ethics can keep up.

 

[annabenedetti]

I wouldn't worry too much about this robot, Anna. The company that created this robot, Boston Dynamics, had been in business for 24+ years. To this day Boston Dynamics has yet to commercialize a single robot. Google purchased them a while back then this past May announced they were selling the company. A major reason for the sale is that Google felt Boston Dynamics was not really close to having a viable commercial product in the near term future. Boston Dynamics is great at creating fancy prototypes and making cool videos but it's quite obvious their prototypes are of little practical use. They have lived off DARPA/DOD money for years and it's quite telling that the US Military has never made any operational use of their robots.

Thanks for that information, TB. I know that you as well as others here are in related fields and know more about these things. Correct me if you don't agree, but I do think that you can compare those robots to the brick cell phones of the 80s and can pretty much see where the exponential leaps and bounds of AI might take us in another 30 years.

Plus, I'm sure the military is much farther along in their own labs and too highly classified to talked about on random tech pages and such.

 

[annabenedetti]

I think people have been watching way too many sci-fi movies. We are a long, long way from having Terminator style robot armies, if ever.

Maybe a long way, but if ever? Did anyone who had an early 80s cell phone ever imagine the power of what they'd be holding in their hand in 2016? It would have sounded like something from The Jetsons. And yet, here we are. Who knows what will be possible in 2040?

 

[rexlunae]

Maybe a long way, but if ever? Did anyone who had an early 80s cell phone ever imagine the power of what they'd be holding in their hand in 2016? It would have sounded like something from The Jetsons. And yet, here we are. Who knows what will be possible in 2040?

My robotic vacuum is a lot smaller than Rosie.

 

[Quincy]

The bots you find in a computer game are barely or not-at-all real AIs. They're usually programmed with pretty simple behavior. A real AI learns and adapts like a real intelligence.

Well, they're basically hacks, but still, aren't even robots which play chess or solve rubik's cubes just pulling solutions from whatever predictions are programmed into them? What I mean is, they don't learn by problem solving in an organic sense. There's no emergent knowledge or experience gained, right? They're still limited by the program. I don't buy into the idea that one day there will be a.i. so advanced it can outperform humans. That's like that house of the future dreams from the 1950s. The Jetsons stuff.

You don't really program an AI. Of course, you have to write some code, but mostly you train an AI.

Isn't it limited by the training, though?

 

[The Berean]

Maybe a long way, but if ever? Did anyone who had an early 80s cell phone ever imagine the power of what they'd be holding in their hand in 2016? It would have sounded like something from The Jetsons. And yet, here we are. Who knows what will be possible in 2040?

Obviously, ever is a long time. The leap from the brick cell phones to today's smart phones was largely driven by the miniaturization of electronics. Terminator style robots are several orders of magnitude more complex than smart phones IMO. I look at it this way. As a teen in the early 1980's I remember reading about how nuclear fusion, flying cars, and Rosie the Robot were just around the corner, maybe 10-15 years away. Well, here we are 35 years later and still no nuclear fusion power plants, no practical commercialized flying cars, and no Rosie the Robots yet. The problem with walking robots is not the AI but the mechanisms and mechanics. Making a bipedal machine with the agility and running ability of a trained soldier is extremely difficult to do from an mechanical engineering standpoint. It's not impossible of course but it is much slower than most people realize.

Sent from my SM-G920V using TheologyOnline mobile app

 

[Quincy]

Obviously, ever is a long time. The leap from the brick cell phones to today's smart phones was largely driven by the miniaturization of electronics. Terminator style robots are several orders of magnitude more complex than smart phones IMO. I look at it this way. As a teen in the early 1980's I remember reading about how nuclear fusion, flying cars, and Rosie the Robot were just around the corner, maybe 10-15 years away. Well, here we are 35 years later and still no nuclear fusion power plants, no practical commercialized flying cars, and no Rosie the Robots yet. The problem with walking robots is not the AI but the mechanisms and mechanics. Making a bipedal machine with the agility and running ability of a trained soldier is extremely difficult to do from an mechanical engineering standpoint. It's not impossible of course but it is much slower than most people realize.

Sent from my SM-G920V using TheologyOnline mobile app

I'd hate to be the wandering nomad in the middle-east, first to witness a galloping metallic gazelle running by camp on a scouting mission lain:

 

[annabenedetti]

Obviously, ever is a long time. The leap from the brick cell phones to today's smart phones was largely driven by the miniaturization of electronics. Terminator style robots are several orders of magnitude more complex than smart phones IMO. I look at it this way. As a teen in the early 1980's I remember reading about how nuclear fusion, flying cars, and Rosie the Robot were just around the corner, maybe 10-15 years away. Well, here we are 35 years later and still no nuclear fusion power plants, no practical commercialized flying cars, and no Rosie the Robots yet. The problem with walking robots is not the AI but the mechanisms and mechanics. Making a bipedal machine with the agility and running ability of a trained soldier is extremely difficult to do from an mechanical engineering standpoint. It's not impossible of course but it is much slower than most people realize.

What about a robot that's not bipedal? What about something that moves like a machine, but thinks like a human? A shape we haven't seen yet, not one that seeks to look and move like a human?

Not contesting, just thinking.

 

[annabenedetti]

I'd hate to be the wandering nomad in the middle-east, first to witness a galloping metallic gazelle running by camp on a scouting mission lain:

Wow, Quincy. That's just so visual. And a scary thought.