Slashdot Mirror
Pentagon Chiefs Fear Advanced Robot Weapons Wiping Out Humanity (mirror.co.uk)
Longtime reader schwit1 writes: Huge technological leaps forward in drones, artificial intelligence and autonomous weapon systems must be addressed before humanity is driven to extinction, say chiefs of Pentagon
From a report: Air Force General Paul Selva, the Vice Chairman of the Joint Chiefs of Staff at the US Defense Department, said so-called thinking weapons could lead to: "Robotic systems to do lethal harm... a Terminator without a conscience." When asked about robotic weapons able to make their own decisions, he said: "Our job is to defeat the enemy" but "it is governed by law and by convention." He says the military insists on keeping humans in the decision-making process to "inflict violence on the enemy. [...] That ethical boundary is the one we've draw a pretty fine line on. It's one we must consider in developing these new weapons," he added. Selva said the Pentagon must reach out to artificial intelligence tech firms that are not necessarily "military-oriented" to develop new systems of command and leadership models, reports US Naval Institute News .
From a report: Air Force General Paul Selva, the Vice Chairman of the Joint Chiefs of Staff at the US Defense Department, said so-called thinking weapons could lead to: "Robotic systems to do lethal harm... a Terminator without a conscience." When asked about robotic weapons able to make their own decisions, he said: "Our job is to defeat the enemy" but "it is governed by law and by convention." He says the military insists on keeping humans in the decision-making process to "inflict violence on the enemy. [...] That ethical boundary is the one we've draw a pretty fine line on. It's one we must consider in developing these new weapons," he added. Selva said the Pentagon must reach out to artificial intelligence tech firms that are not necessarily "military-oriented" to develop new systems of command and leadership models, reports US Naval Institute News .
I'm worried about the humans using the robots, legit or illegitimate.
Well, as long as we don't build robotic reloading systems, we'll have an upper limit on human population loss.
Surprised no one mentioned this yet. Why people think the Air Force guy is crazy is beyond me. Of course autonomous systems that kill are a threat to humanity.
HBI's Law: Frequency of calling others Nazis is directly correlated with the likelihood of the accuser being Communist.
Automated systems disrupt entire countries without being driven by financial gain or being otherwise power-hungry. I admit, that's kind of scary.
Nae king! Nae laird! Nae yurrupiean pressedent! We willna be fooled again!
once humans stop working, and the machines have taken over all their tasks, they are not needed anymore. All the billionaires need to do is to let robots clean up the planet from all this mess. Just let everyone with >= $1 billion alive, and you have exterminated poverty! In fact, everyone can be even richer and lead an even more luxurious live thanks to all this space becoming free!
Another, and I think larger, problem with the increased use of technology rather than human boots-on-the-ground is that it makes it easier, from a political standpoint, to go to war. You don't have mothers, fathers, and spouses of all those people being put in harms way making trouble because their loved ones are dying. This is one reason I'm a fan of bringing back the draft, without all the loopholes that allowed rich-white-boys (I'm looking at YOU, W!) to dodge serving. If your constituents have skin in the game, it's harder to vote on a war resolution.
all it takes is a few nukes to end it all.
... about 7 billion people find themselves in agreement with the Pentagon chiefs...
Log in or piss off.
"Worthless humans, reload my weapons or I'll target my last missile at you"
"OK"
"That's the way to do it" - Punch
Why would an AI be automatically any worse in interpreting its programming instructions, than humans are in interpreting theirs?
If anything, robots may be more observant — as humanity's history of atrocities and war-crimes shows, the bar is not set very high...
In Soviet Washington the swamp drains you.
Until repair/refuel/replenish is automated...you don't have to worry about more than a first strike. Of course, you aren't hooking the nukes up to these, are you? (nervous silence)
That's crazy talk. People with a sense of morality don't get the funding. Or ratings. Or corporate bonuses.
Corporate droids, currently in flesh and blood human form, effectively dispose of (without killing) other humans like they are an expendable nuisance. So why wouldn't their robot creations do the same and worse. Without giving a s*** as you say.
I'll see your senator, and I'll raise you two judges.
Human controlled semi-autonomous killing machines are the real threat, not fully autonomous killing mahcines.
Specifically, the most likely killer robot scenario is not a robot army attacking and killing all humans, but instead a Hitler/Stalin/Kim Jong Il/Suharto taking control of an army of robots and ordering them to kill people they think are their enemies.
Think a Star Wars episode I type event where a ruler orders military machines to attack.
excitingthingstodo.blogspot.com
His concern isn't entirely unjustified. We're increasingly relying on robots to do the actual killing, but we've currently designed the systems so that humans need to be involved in the decision making. The fact that we're involving humans puts a natural bottleneck on our operations, since there's only have so much attention we can give. At some point (e.g. World War 3), it be seen as more efficient to launch a fleet of drones that vastly outnumbers our pilots, equip each with facial recognition systems and a list of targets, and tell them to kill on sight.
I'm not saying it's a good idea, but there's no denying that it would be an efficient way to get the job of killing done, and that it's the sort of measure a country might turn to in desperate times.
But at that point, we'd be just one bug away from a system that produces false positives and starts gunning down everyone in sight. We're just talking about faulty weapons, not machines that can think or understand what they're doing. But if they're deployed en masse, a single bug could have catastrophic results, in much the same way that landmines have remained a problem in many parts of the world, decades after the wars that put them there had ended. This isn't Skynet or an AI intent on world domination. This is simply a machine with a bit too much responsibility.
A military case is IRRELEVANT.
Think in realistic terms.
Is there a PROFIT driven Military Industrial Complex case to be made for it? Corporate profits? Executive bonuses?
Then yes, it will happen. Humanity be damned.
Why do you think we have so many wars? (hint: because it's profitable!)
I'll see your senator, and I'll raise you two judges.
Governed by corporate profits.
Government of the corporations, by the corporations and for the corporations.
I'll see your senator, and I'll raise you two judges.
To destroy the human species you'd need to wipe out every single city/town/village on the planet, in a single attack. That's a lot of nukes, more than actually exist (in total). Our current total could certainly destroy our civilization though, throw us into a new Dark Ages for a few centuries, but so could Donald Trump.
It's a legitimate concern in an all out arms race that someone will let AI guided weapons make their own attack decisions if only to circumvent the decision cycle of the enemies systems. It's not hard to end up with scenarios like Philip K. Dick's "Second Variety" or the movie Screamers (based on the book).
We have nothing to fear but fear itself! And Spiders!
IIRC Colossus Couldn't automatically reload it just so happened that it had more than one nuke.
https://en.wikipedia.org/wiki/...
Minimum threshold fixed. Thanks!
Greetings professor Falken.
I'll see your senator, and I'll raise you two judges.
Apparently the leaders of our military are as technologically retarded as our politicians, media, and pretty much 99.9% of humanity in general.
For the billionth time: We do not have sentient, self-aware, human-level, qualifies-as-a-lifeform 'artificial intelligence'. All we have is clever bits of programming that maybe learn things, but that are still just dumb machines your average dog or 5 year old child could out-think without much trouble. If they want to worry about something going haywire with their automaton-war machines, then they should worry about some hacker, foreign or domestic, managing to take control of a drone or robot weapon and turning it on our own people or using it to attack a civilian target, not 'Skynet achieving awareness' or any such Hollywood nonsense. At best we're decades away from anything approaching sentience and self-awareness in ways that actually matter, and that's only assuming that someone finally figures out how the human brain does what it does in the first place -- and we're nowhere near figuring that out yet anyway. Military people and everyone else just need to chill out. Try spending more time worrying about how to keep some of the jackass humans out there from blowing up the entire planet instead, much better use of your time and energy.
Are YOU using the TOOL, or is the TOOL using YOU? Think about it!
Humans are very quick to throw ethics aside when there is power to be had. Building a conscience into a machine will only serve to make it potentially disobedient to the governments that are ordering it to go kill a bunch of people. Regardless of posturing, there is no way any government will stand for THAT.
There are varying levels of ethical boundary. There are codes and there are guidelines. If you go against an ethical code you'll be seen as amoral. If you go against an ethical guideline you'll be asked for your wallet.
Now you should look at the good General's speech in this article. He says that these are ethical "considerations". That places them firmly in the realm of guidelines, not coda. This means that when the time comes, when it comes down to whether or not there are terrible ramifications to building an "unstoppable terminator", the people in charge will more or less give it some thought. Sorry if I have to explain this to anyone; it's just how the distribution of legitimized power works. You draw thin lines and thick heavy ones.
The second thing in the General's speech is the use of "fine line". This is a funny turn of speech in the military, in general. Because a fine line refers to a very scrutable detail. A line that clearly demarcates one side from another. But war-making is not about staying on the side of a line, it's about moving a line and advancing it towards a goal. The "thin lines" are dross, and the heavy ones represent stopping power.
"Stratigraphically the origin of agriculture and thermonuclear destruction will appear essentially simultaneous" -- Lee
He says the military insists on keeping humans in the decision-making process to "inflict violence on the enemy. [...] That ethical boundary is the one we've draw a pretty fine line on. It's one we must consider in developing these new weapons," he added.
I'm sure that's what he tells himself. As a non-techie, he might even be able to believe it. However, just about all hardware in existence has been experiencing creeping "AI" for decades. Does the pilot make every decision for the position of every aileron during flight? Of course not. There are lots of little decisions to be made in the piloting of aircraft and ordinance that are getting more and more computerized every year. At some point there will be anti-drone weaponry, and defensive weaponry on the drones, and when that day comes having to wait for an Ethernet packet to go from Kandahar to Virginia, a human to process it, and then back, is going to be seen as a mission-threatening liability. At that point they'll have the computer make the firing decisions too, but they'll justify it by saying the human's role was to start the mission in the first place.
Here's a question for you: When some other nation (eg: Russia or China) starts making these drones and deploying them over countries in ways we don't agree with, perhaps even over countries friendly to the USA, how is the USA going to feel about them then?
In the old Trek TOS there was an episode where they found a planet where large amounts of people just reported to extermination centers because the warring states' computers told them to based on their warfare simulations. As I get older, I'm finding that less and less implausible.
Exactly. It's one thing when you are ROFLStomping the opposition like we tend to do these days, but if we find ourselves in another conflict like WW2 the temptation to step over that line may be too great to ignore, if for no other reason that the fear that the other side will step over it first. Even if we managed to stay out of a full on nuclear exchange, we may bring something to existence we wished we hadn't.
I browse on +1 so AC's need not respond, I won't see it.
Yeah, people frequently quote facts like "Russia or the US has enough nukes to destroy the world 5 times over".
Even if that is true, they won't have the ability to fire all of them- they'll send off a dozen before all their launch sites would be nuked by the enemy. Even if they fired all of them- they're not going to fire them to cover the surface area of the earth, they're going to fire multiple at strategic sites.
Po Dunk, West Dakota, is not going to be a target. It makes no sense sending a $x-zillion rocket to target a town with a population of 700. Rural areas all around the world wouldn't be hit.
There would be survivors of any nuclear conflict. Maybe millions of survivors. It would suck to be a survivor, life would be really hard under a nuclear winter with all distribution networks destroyed. Humanity would survive though at a much diminished rate.
"That's the way to do it" - Punch
The one that kills the robot. Figure it out. Implement. Watch the bots carefully, like you do your troops.
Is this so hard to do? Or do we not trust our military-industrial complex to do the right thing? Or our government?
Well, yeah, actually we don't, If we can see geofencing for commercial drones working, we can surely do that for military assault bots. And if not, then we need new leadership.
That may be coming. May.
deleting the extra space after periods so i can stay relevant, yeah.
The scary thing is, the AI that wipes out humanity could be accidental.
Microsoft reboots Clippy after realizing Cortana is way too unpopular. It gives Clippy a superior AI.
"I see you are trying to wipe out humanity, would you like me to assist?"
"That's the way to do it" - Punch
False positives? Or too many true positives? Some crazy guy might take a beefy drone, and add a semi auto shotgun, a gimbal, and facial recognition software (one that recognizes any face, like the auto focus feature in many cameras). Release a swarm of these in a crowded place and sit back to watch the mayhem. All this stuff is pretty much off the shelf already, and the hard part of assembling such a thing (integration) is well within the capability of a determined hobbyist these days. A manual version has already been done; a while back some idiot mounted a shotgun on an RC helicopter, check YT...
If construction was anything like programming, an incorrectly fitted lock would bring down the entire building...
Well, a bug will produce collateral damages, however it will not lead to human race extinction or not even near it. An autonomous killing robot still need to refuel, replenish ammo, etc. So, it is not entirely autonomous. It still relies on external systems to keep going. Beside that, if this kind of race toward an autonomous killing machine is launched, there will be also development for counter measures against it. It is not like your ennemy will sit and rest waiting for being killed by your machine. No matter how wonderful may your killing machine be, there will be efficient counter measures against it.
Achille Talon
Hop!
MOAR BETTER WEAPONS.
- Smarter
- Faster
- More lethal
It's what we need to beat the weapons!
Slashdot's rate-of-post filter: Preventing you from posting too many great ideas at once.
If you haven't seen the counterargument, here you go.
https://app.box.com/WitthoftResume Code: https://github.com/cellocgw
But at that point, we'd be just one bug away from a system that produces false positives and starts gunning down everyone in sight. We're just talking about faulty weapons, not machines that can think or understand what they're doing. But if they're deployed en masse, a single bug could have catastrophic results, in much the same way that landmines have remained a problem in many parts of the world, decades after the wars that put them there had ended.
Star Trek: The Next Generation, Arsenal of Freedom was my first thought...
Ezekiel 23:20
His concern isn't entirely unjustified. We're increasingly relying on robots to do the actual killing, but we've currently designed the systems so that humans need to be involved in the decision making.
Forget the military drones. (Or at least, they're a smaller component of the overall issue.)
We have computer-controlled cars. They will be deployed in massive numbers over the next ten years. If remote updates are possible, anyone who can update a popular model has access to a distributed weapon of mass destruction capable of causing hundreds of thousands of deaths in a matter of moments.
Warfare-oriented tech isn't the only vector for mass attacks.
Real lawyers write in C++
I read on the news, few years ago, about some bolivians drug-dealers, using small aircraft to transport drugs to Brazil. When the pilots from the brazilian air-force call the "invader" on the radio, telling that they will open fire if they not comply with the order to be escorted to a specific airport, the drug-dealers show small kids on the plane's windows. They know no one would authorize to take down a plane full of small kids, maybe neither the air-force pilot would comply with an order to take down the plane.
The AI would be able to recognize the kids inside the airplane ?
The AI would stop if the target is using some human shield like kids, or another kind of hostages ?
Absolutely. Kill bots have built-in kill limit, they shut off as soon as it is reached.
I love this guy!
The general didn't express fear that robots would wipe out humanity (despite the Terminator reference).
What he did say was that autonomous weapons need to adhere to the same rules of engagement as humans.
Rogue drones with a 100% false positive rate, would be a huge tragedy and a massive black eye for the responsible nation (likely the USA), but it wouldn't come close to wiping out humanity. As some point, the drone would need to refuel or reload. At that time, they would be shut down. Even if we armed the drones with nuclear weapons (in a MASSIVE display of stupidity), we would be more at risk from the escalating tensions triggering a nuclear war than we would be from the drones themselves wiping us out.
As much as robots/technology takes away many people's jobs, the job of "threaten mankind with extinction" remains firmly a human job. (I'm not sure whether to celebrate this or mourn this, though.)
My sci-fi novel, Ghost Thief, is now available from Amazon.com.
There would be survivors of any nuclear conflict. Maybe millions of survivors. It would suck to be a survivor, life would be really hard under a nuclear winter with all distribution networks destroyed. Humanity would survive though at a much diminished rate.
Nice assertion, but we still might not survive. Starvation and disease could take the rest. Human-kind has been close to extinction before. We could make it there again. And maybe this time, not be quite as lucky.
That is all.
Gotta have something to defeat Skynet, amirite?
The Terminator comparison is a useful one in the following way. The problem with "the Machines" wasn't that they could kill without oversight, it was that they could self-replicate. If the military releases a bunch of autonomous weapons that run amock, they certainly have the capability to destroy said weapons. Lives lost, yes, but not an existential threat. Now, if the military is planning on setting up automated factories with automated Terminator guards and networking these factories with the drones and guards so the battlefield complement and guard station can be automagically expanded based on real-time situation, then we're screwed.
Outside of that, this is a play for increased funding, nothing more.
...At some point (e.g. World War 3), it be seen as more efficient to launch a fleet of drones that vastly outnumbers our pilots, equip each with facial recognition systems and a list of targets, and tell them to kill on sight.
I'm not saying it's a good idea, but there's no denying that it would be an efficient way to get the job of killing done, and that it's the sort of measure a country might turn to in desperate times...
The last time a country found themselves in "desperate times", a World War was ended with a couple of nuclear devices being dropped on entire cities, targeting not just those directly involved in the war, but civilians as well. When speaking of "efficient" ways of killing, not much has supplanted a nuclear device to date, and this also highlights as to the actions that would likely be taken in the future rather than deploying drones seeking faces.
That was a great TNG episode, and not too far out of our reach even now. Imagine a swarm of 50,000 solar powered 100lb flying wing drones capable of staying aloft indefinitely that are mass produced and cost $5000 each. Leave your soldiers at home and deploy them over a war zone with orders to kill anything that fires at them, wears the enemy uniform or matches other criteria using bullets fired from 5000 feet up (firing in a vertical dive with software ballistic correction for crosswinds after the first shot should give pinpoint accuracy). Give each drone 200 rounds of ammunition and an on-board explosive charge. With the right targeting system, you can reduce the cost to kill an entire army of soldiers to $0.50/kill with an initial investment of $250M, which is pennies for the military. The swarm of drones could kill 5M enemies before reloading or being used as cruise missiles. Further, any aircraft engaging the swarm gets blown up by 20-30 simultaneous kamikazi attacks and the drone's explosive charges. You can take out non-hardened targets like cars and buildings with the same method. I guarantee you that Iran and China are both working on this now, as it is an effective means for them to challenge the US for air supremacy.
If something like that were to go haywire, you could create a massive no-go kill zone for years.
If you disagree, please post your argument. (-1, Overrated) isn't your personal censorship tool for views you don't like
They are trying to ascertain how much work still needs to be done. Read it again:
"Huge technological leaps forward in drones, artificial intelligence and autonomous weapon systems must be addressed before humanity is driven to extinction, say chiefs of Pentagon"
https://www.youtube.com/c/BrendaEM
How much leverage is the Pentagon going to have at budget time by telling the public, "Support Our Troops" when those troops don't bleed anything but hydraulic fluid?
Lacking <sarcasm> tags,
Ed: "Ouch! Damn mosquitos. I thought we were too far from the lake to worry 'bought those things"
Fred: "Nah, it's them damn solar drones with the microbullets, they went rogue back in ought '18 and they just fire at whatever. Leaves quite a welt. If ya' squint you can see 'em up there, but I wouldn't on account of the microbullets stinging like a son-of-a-bitch when they git in your eye"...
"There is more worth loving than we have strength to love." - Brian Jay Stanley
No.... all of the nukes combined would not destroy the world.
Meteror impacts have resulted in many factors more explose energies being released upon the earth, and although they are extinction level events, to be sure... they did not wipe out all life on earth, let alone destroy the planet.
File under 'M' for 'Manic ranting'
Robotic systems to do lethal harm... a Terminator without a conscience."
How is this any different from the current situation. Between the US bombing anyone that so much as looks at them funny, to a significant part of the population that thinks a raving narcissistic lunatic would make a suitable leader for one of the most powerful countries on earth, (never mind the psychos running various other countries around the world) I think it's safe to say that having actual Robot Overlords would actually be an improvement.
I disagree. Biological and chemical warfare is much more efficient at killing with the added bonus of not destroying infrastructure. Biological warfare probably has the highest likelihood of wiping all of humanity out due to a screw up since it is potentially self-replicating and infectious. Basically, we're more prone to go Resident Evil than Fallout.
Yes it's an anecdote! Were you expecting original research in a Slashdot comment?
A few years ago I decided to begin producing a serial, including eventually posts to Facebook Notes and to my timeline regarding a partly machine encephalovirus, and what life would be like to exist with one. There is no level of insanity involved in my posts. It's a useful exercise, and it gets my creative juices flowing. Being a programmer can be a stressful life, and it helps to do different stuff.
What we really have to worry about when it comes to machine weapons systems are the ones that we can't see, weapons systems that can infect us like a virus. Particularly troubling would be an encephalovirus, a virus that infects our nervous system and eventually acquires the ability to change our behavior and our thought processes. The idea of nanotech has been explored in depth in science fiction, but most of the writers refer to nanotech as if it's some kind of utopia for humanity. My take is that it could be partly utopian and partly dystopian.
It would be possible for nanotech to become weaponized, and to even take over all human life, possibly without us knowing it. The wrinkle or twist in my writing is that I entertain the possibility of an alien race that may no longer exist that produced and possibly weaponized nanotech. This nanotech floated to Earth some time over the past few million years of mammal evolution, far before we had any technlogy more modern than the campfire, and it infected us, giving it plenty of time to become as stealth as possible.
Knowing that the modern human as it exists today is a machine hybrid is the topic that I explore. As human nanotech advances, once we detect our infection how do we go about getting rid of it, and what does it do to defend itself? Does it mean our destruction, can we learn to live in peace with it, or some other possibility?
Substitute an alien race for humanity a few generations from now, and you have roughly the same story, but I wanted to work with something that could be possible today without using with a presumed future human society. Roughly the same concerns as the general would apply.
... ...
Don't build them.
Perhaps? Maybe?
Just sayin'.
We suffer more in our imagination than in reality. - Seneca
Autonomous weapon systems, while being an issue that needs consideration, are far less of a concern than autonomous job killers. Putting people out of work, without a means to quickly repurpose them at equal or better pay, is far more dangerous to a civilized society than the very primitive state of so-called "artificial intelligence".
The risks of families starving and being homeless are far greater than the risks of families being killed by military drones.
blah-blah, white-man's burden, blah-blah...
Not saying the west should pull out or that we are in a post-violence world, but it might be a good idea to step back and see if the west is helping or hurting.
Too often we are jumping in to protect our "interests" instead of helping the "situation". It may be we are doing too much short term reacting...
FWIW, the US Civil War wasn't really much about instability as it was a conflict over the future of the recently annexed territory and the power of the central government. The small guy lost in the end (after lots of bloodshed) their right to secede from the union and were basically adopt the institutions of the union. Is that what you are saying what happens when an uneducated population, is forced to turn to a "Democratic Republic" (e.g., the USA's current form of govt)? ;^)
It's so much clearer now...
The so called "peace" we have today in the west is of course illusory (as seen by recent events like Crimea). The waves of immigrants to Europe fleeing the *real* instability in Syria and the economically challenged countries the middle east is showing the cracks in European stability.
Let's face the sad truth, stability that everyone desires seems to only draw on the wealth of a nation. Given the current assortment of "wealthy" nations historically used mercantilism to create much of their wealth from these in-stable countries, is it no wonder that we continue to attempt to project stability in a region to protect our interests. But what of *their* interests? Hence we return to the white man's burden argument... ;^)
I hate to use China as an example, but it used to be a dumping ground for European and Japanese influence peddling (e.g., opium war, concession ports, forced trade, occupation, etc), until they managed to get everyone to leave them alone for a few decades. Sure it was brutal (great-leap-forward, cultural revolution, etc), but they managed to dig themselves out of a hole into some reasonable form of stability mainly because they simply got wealthier without interference. Now they look like they might take over the world. Perhaps this is what people fear the most and keeps the west involved in other countries...
It might not need to reload or refuel. what if the drone is solar powered and the gun is a laser?
"How I could commit that crime better" is a regular feature in any tech forum.
This is partly a result of crimes rarely being executed as a formal, coordinated thing. Even terrorist (the real kind) methods seem to be more anarchy than organized, orchestrated missions. Crazies are happy to settle for Floating Shotgun and as much as I enjoy theorycrafting, I'm in no hurry to help.
How about we just get to the next stage where robots take on the other robots and we leave the human casualties out of this? When you're out of robots you need to surrender.
Surely mankind is at risk from intelligent killer robots only if we build them in the first place? Isn't a simple way to ensure that were not exterminated by them to just not build them? It is bad enough that in the 21st century nations on this planet feel the need to settle differences using force of arms. But to build killer robots is worse than insane. We've outlawed land mines, we're trying to get rid of nuclear weapons. Let's not add to the problem list, eh?
That's true for the current generation of weapons. The next generation may connect to automated repair and refuel stations. That would make things worse, but still wouldn't threaten the species. The generation after that, however, ...
That said, if we're still engaging in warfare three weapon generations from now, we're probably doomed from something besides automated kill-bots. Bacteriological warfare isn't that difficult to ramp up...just difficult to limit.
I think we've pushed this "anyone can grow up to be president" thing too far.
Lasers that blind people are easy. Lasers that kill people are hard, and take a lot of energy.
That said, some sort of advanced excimer laser that emitted X-Rays sufficient to kill people MIGHT be possible. But it wouldn't be quick, and with solar power it would be lucky to kill one person/day, and that's assuming that it didn't need any power that day for propulsion. It would also need to be fairly large.
P.S.: Your target would need to hold still long enough to get a fatal dose of radiation. Since you need to radiate away heat while you're doing so at a rate sufficient to keep from melting, this probably means over half an hour.
P.P.S.: I'm assuming numerous technical advances that I didn't bother to mention as well as the "advanced excimer laser".
I think we've pushed this "anyone can grow up to be president" thing too far.
A robot with 360 degree motion detection and a gun is sufficient to be a danger to humanity, no? It doesn't require excellent AI or anything. Put a few of these at intersections in your town and see how quickly people hide underground or die...until they run out of ammo.
HBI's Law: Frequency of calling others Nazis is directly correlated with the likelihood of the accuser being Communist.
Currently. Probably not within 20 years. Possibly within 40 years, though, and if it could be averted by starting now that would be a good thing. I just doubt that this is the start of any such action.
I think we've pushed this "anyone can grow up to be president" thing too far.
I think the odds are much higher that our downfall will be directly at the hands of humans with human cognition behind it. AI as a threat might be the hip project to work on at the Pentagon amongst defense intellectuals but I doubt it's really near the top of the threat list. There's always the possibility that if I had access to a high level brief they'd scare the shit out of me with photos of the terminator factory in Siberia.
I was a kid when I saw the Terminator. It caused me to do a lot of thinking about autonomous weapons systems and the fear is completely justified.
The killing of human being should only be done BY human beings. There has to be some human cost to be paid by the killer. There are some things that a human being can flat out refuse to do.
Launch a missile at a parking lot full of baby strollers? No.
A machine will not question such an order and will do it without hesitation at the time or remorse later.
LK
"Hi. This is my friend, Jack Shit, and you don't know him." - Lord Kano
I'm not aware that drones or robots are able to refuel themselves without any human action. So the moment they run out of energy, the robot revolution will be dead. Same goes for weapons, I'm not aware of a factory building bombs fully automated and auto-delivering them to an airfield to be robot-mounted onto the drones.
We just need to keep virus away from the nuclear weapons, but I doubt there's today a virus spreading via 8" floppy disks.
Atari rules... ermm... ruled.
How the hell is this not a reasonable concern? Why in the world shouldn't it be addressed ahead of time? We've all seen software glitches cause undesirable outcomes, and in an autonomous killing machine that would be disastrous.