Slashdot Mirror
Top US General Warns Against Rogue Killer Robots (thehill.com)
Long-time Slashdot reader Zorro quotes The Hill:
The second-highest-ranking general in the U.S. military last Tuesday warned lawmakers against equipping the armed forces with autonomous weapons systems... Gen. Paul Selva warned lawmakers that the military should keep "the ethical rules of war in place lest we unleash on humanity a set of robots that we don't know how to control. I don't think it's reasonable for us to put robots in charge of whether or not we take a human life," Selva told the committee.
There's already a Defense Department directive that requires humans in the decision-making process for lethal autonomous weapons systems. But it expires later this year...
There's already a Defense Department directive that requires humans in the decision-making process for lethal autonomous weapons systems. But it expires later this year...
If war was ethical, only leaders would fight.
Top US General concerned about future job security. Worries the human element will soon not be a requirement when it comes to warfare.
This is a big deal in a country where War and Combat are glorified and have seeped into the facets of everyday life.
Our bigger problem at the moment are killer generals (US and elsewhere).
"Killer robots" are going to be created. As it gets easier and easier to do with off-the-shelf and/or printed components it is inevitable. Once that happens what comes next will be a matter of cyber security and cyber warfare. The "winner" in any war that uses autonomous killing machines as combatants will be the side with the best electronic warfare systems.
Gen. Paul Selva probably understands that this is currently not his government, and recent administrations either have not gotten the memo or are playing their cards very close to their chest. I suspect he is much more worried about creating efficient killing machines that get co-opted and controlled by his adversaries than some AI going rogue and asserting their position atop Earth's food chain.
I am not interested in articles about life extension advancements.
Indeed, I'd personally be more worried about how they solve a problem of the people in power being capable of simply ordering the robots to kill everyone and robots not going rogue and following instructions to the letter.
I love the modern idea that works of fiction, specifically written to advance a particular point of view, are somehow indicative of how reality works. It's a movie, it's entertainment.
Shutting down free speech with violence isn't fighting fascism. It IS fascism!
If you're a ST:TNG fan, think of the episode "Arsenal of Freedom".
Autonomous weapons sold to both sides of a planet at war, both populations killed by those weapons, all that was left were the autonomous weapons.
when it comes to killing innocent people (for political gains and self-interest). American young men are willing as ever to enlist to "get some" and to become a "hero" in Afghanistan or wherever. And, as we've seen, they have little consideration for innocent lives.
Why would they suddenly think putting a gun in the hands of a robot would be a bad idea? A trick to make them seem considerate, I guess.
I love the modern idea that works of fiction, specifically written to advance a particular point of view, are somehow indicative of how reality works. It's a movie, it's entertainment.
Yeah, in reality the AI wouldn't be a rogue one but being a good little german and following the orders to the letter when it exterminates the starving rioting unemployed serfs.
The solution to the intelligent robot problem may be to do what the robot designers did in the Star Wars universe -- program personality disorders into the intelligent robots.
Just because you are paranoid does not mean that no-one is out to get you.
First came the R/C devices used, then the semi-autonomous devices with minimal ability to deviate from a pre-planned route. Then came the more-or-less fully autonomous devices, where you give it a map and a target and a 'go' order. Right now, the machines return video or other data and a human gives the final OK.
That's fine (for the US and allies) while they're the only ones who can deploy that level of tech in the field, but as everyone else catches up, it'll be the ones that take the humans out of the loop that respond faster and win the engagements. And the US won't sit by and watch as that happens, they'll remove their human oversight.
The next step will be false flag ops, blaming the enemy's bad software. And, eventually, there will be a bad map update or a malicious instruction and you'll have a drone swarm committing genocide for you.
This is inevitable, and rather than try and prevent it (which is futile) we ought to be worrying more about counter-strategies. Maybe we need to say that we can't be as free as we'd like to be, and drones have to go - that anything over a certain size (big enough to carry a dangerous payload a significant distance) will be shot down on sight unless it's a registered, transponder-carrying device.
I can honestly see the day when densely populated areas are protected by automated anti-drone systems. It's just too easy to launch hundreds of moderate-sized devices at an urban center to sow chaos and fear.
And just wait until the first self-driving car bomb...
The military doesn't have the luxury of holding back because of the worry about all the negative consequences of new military technology. If the technology can exist, someone will develop it. The best defense I can think of is developing it yourself so at least you can understand the true dangers and potentially build countermeasures against them.
Yet we use this 'luxury' when it comes to many types of existing weapons. And what choice does humanity have? We are well beyond local tribes with spears and shields. The western minority powers can literally make everybody on the planet extinct if they want. If we must just accept that there is no way to build lasting peace, then we are simply counting down to our own extinction as every generation of smartphone gets better at ordering pizza and looking up trivia.
The thing that scares me the most about these weapons however, is that it removes the democratic element of war. to fight a war you need a powerful army but also a loyal army. That same mass of armed civilians can turn against a ruler if they lose their populous appeal. This is why countries like north korean must run massive propaganda campaigns, and why much of the key to the rise of fascism was its ability to use new forms of mass media. It is why a free press and education are seen as essential elements in the fight against a repeat of humanity's past atrocities.
But once you have autonomous armies, you no longer need trained civilians. A government can indeed use that army to control citizens and ensure it remains in power against majority rule. The political implications of this should scare anyone - we have never really had such a threat before. For me this threat from within is far greater than the meaningless risk of open conflict between nuclear armed states.
While the notion of a Terminator-sequel, self-aware system like Skynet is still firmly rooted in the realm of science fiction, the idea that we can make an autonomous system designed to kill humans isn't neatly so far-fetched, nor is the idea that some of them may go beyond their expected parameters. They'd be dumb killing machines, little more than a modern, mobile version of mines, but they'd be more than capable of killing people who happened to wander into their path until they ran out of ammo.
They're unlikely to be existential threats to humanity as a whole unless we give them incredibly lethal payloads and the means to re-equip themselves autonomously, but so long as these dumb machines still depend on us for ammo, we don't need to worry about them learning how to equip themselves otherwise, so the threat will always be limited.
autonomous robots will drastically increase the danger of a rogue general, they'll obey orders no matter what
They also increase the danger of a rogue government.
We used to have a Bill of Rights. Now, with the rights gone, all we have left is the bill.
Remember that terminators that can kill people can serve tea as well. The rich and powerful will control these resources and not need that many other people. Every time I read an article on UBI - Universal Basic Income, I think it is more likely to get UBG - Universal Basic Genocide.
I only look human.
My mother is a halfling and my dad is an ogre, so that makes me an Ogreling
Back in the 1960s, the USAF deployed a Surface to air missile called the Bomarc ( https://en.wikipedia.org/wiki/... ) The thing had a range of around 400km and conceptually, could be used to intercept flights of long range bombers headed toward the US. The problem was that the bomarc could have a nuclear warhead. Fine if you want to take out a squadron of bombers someplace out over the Atlantic. But what if you wanted to call off an intercept for some reason? You can tell an F-106 to return to base. But putting a pilotless missile with a nuclear warhead on RTB was considered to be a non-optimum strategy.
I'm not sure the usage was ever resolved. Fortunately or not, the threat switched from long range bombers (which we probably could not actually intercept reliably because of jamming) to ICBMs that we could not intercept reliably because we lacked the technology to intercept them.
The bomarcs were scrapped in the early 1970s.
You can't see ANYTHING from a car, You've got to get out of the goddamned contraption and walk...Edward Abbey
Land mines can be thought of as fully autonomous robots. Perhaps the simplest case of a 'robot'.
Very simple predetermined command to follow: 'When your trigger is tripped, execute your explosion sequence.'
Most nations have banned the use of land mines because of their uncontrolled, autonomous behavior. Once they are set, they stay set and will activate whether tripped by friend or foe.
They will activate when tripped by the little child playing in the field years after the war is over.
The problem the General recognizes in fully autonomous killer robots is the same problem encountered when land mines are used. The robots are just a more complex example.
I agree there's a problem: OTHER nations will make them whether we do or not; and therefore we are forced to pursue similar technology to compete and survive.
Such bots will probably need a relatively simple "kill-switch" mechanism that is independent of the rest of the brain. Thus, if the main brain gets hacked or goes berserk, the independent kill-switch can be contacted to disable the entire thing. Because the kill-switch is (hopefully) a relatively simple mechanism, it's easier to prevent it being hacked. There is a risk the enemy will steal the security codes for the kill-switch, but that's a risk we'll have to accept.
If you think about it, cars and key infrastructure will probably need the same safeguards regardless of whether war-bots are built, because they can face similar problems. It's part of a bigger problem: stopping highly-complex machines that are out of control.
Table-ized A.I.
bullshit, those mines have killed hundreds of rural farmers and maimed thousands. "controlled", ha.
a war with NK will NOT start by a bunch of NK soldiers marching across the DMZ.