QuoteElon Musk and Stephen Hawking Among Hundreds to Urge Ban on Military Robots
By DANIEL VICTORJULY 27, 2015
Elon Musk and Stephen Hawking, along with hundreds of artificial intelligence researchers and experts, are calling for a worldwide ban on so-called autonomous weapons, warning that they could set off a revolution in weaponry comparable to gunpowder and nuclear arms.
In a letter unveiled as researchers gathered at the International Joint Conference on Artificial Intelligence in Buenos Aires on Monday, the signatories argued that the deployment of robots capable of killing while untethered to human operators is "feasible within years, not decades." If development is not cut off, it is only a matter of time before the weapons end up in the hands of terrorists and warlords, they said.
Unlike drones, which require a person to remotely pilot the craft and make targeting decisions, the autonomous weapons would search for and engage targets on their own. Unlike nuclear weapons, they could be made with raw materials that all significant military powers could afford and obtain, making them easier to mass-produce, the authors argued.
The weapons could reduce military casualties by keeping human soldiers off battlefields, but they could also lower the threshold for going to battle, the letter said. "If any major military power pushes ahead with A.I. weapon development, a global arms race is virtually inevitable, and the endpoint of this technological trajectory is obvious: autonomous weapons will become the Kalashnikovs of tomorrow," it said.
Mr. Musk, the head of SpaceX, has raised warnings about artificial intelligence before, calling it probably humanity's "biggest existential threat." Mr. Hawking, the physicist, has written that while development of artificial intelligence could be the biggest event in human history, "Unfortunately, it might also be the last."
The letter said artificial intelligence "has great potential to benefit humanity in many ways." Proponents have predicted applications in fighting disease, mitigating poverty and carrying out rescues. An association with weaponry, though, could set off a backlash that curtails its advancement, the authors said.
Other notable signatories to the letter included Steve Wozniak, the co-founder of Apple; Noam Chomsky, the linguist and political philosopher; and Demis Hassabis, the chief executive of the artificial intelligence company Google DeepMind.
http://www.nytimes.com/2015/07/28/technology/elon-musk-and-stephen-hawking-among-hundreds-to-urge-ban-on-military-robots.html?_r=0
:ph34r:
How exactly do they plan to ensure that individual countries can't obtain the robots? If, say, Russia says it will go ahead with development anyway, what can they do?
Quote from: Monoriu on July 28, 2015, 06:00:03 AM
How exactly do they plan to ensure that individual countries can't obtain the robots? If, say, Russia says it will go ahead with development anyway, what can they do?
Probably with a non-proliferation treaty.
Quote from: Martinus on July 28, 2015, 06:01:21 AM
Probably with a non-proliferation treaty.
And if Russia and China refuse to sign such a treaty?
Then the US makes more, better killer robots. :showoff:
Artificial stupidity worries me far more than artificial intelligence
Musk has done so much for my issues that I can look past this.
Fully-autonomous weapons are already in everyday use. It's the ethics of deliberately using them against humans without a "man in the loop" that's in question. This is known as lethal autonomous robotics. or LARs. Here's an interview I did with a human rights lawyer on the subject.
http://www.army-technology.com/features/featureterminator-on-trial-ethics-lethal-robotics/ (http://www.army-technology.com/features/featureterminator-on-trial-ethics-lethal-robotics/)
Incidentally, the ethics of a machine making a decision about human life is not limited to automated weapons. Automated cars are another example - let's say you are driving in an automated car and you get two (non automated) cars coming from the other side. One is full of kids, who have a high chance of dying if you hit it, but you also get some chance of surviving; the other is a huge truck which will surely kill you. There is no way to avoid hitting either. Are we happy with your automated car making a decision who to kill (including, yourself)?
Tell all those listed in thread to get in line, Ed and I were here first :ultra:
I think Peedy may want to join if it's full-blown Tim-bashing. :hmm:
If Chomsky is against it I'm for it.
I totally agree with them. I can't think of any reason we'd want to create a true AI anyway unless we do in fact want to go extinct. Human minds are pretty awesome.
Speak for yourself.
Brains aren't too bad either
Hmm . . . it's not like humans reliably make such great decisions.
Quote from: The Minsky Moment on July 28, 2015, 11:04:12 AM
Hmm . . . it's not like humans reliably make such great decisions.
We had a discussion about you being a precursor agent. One more data point in support of that conclusion.
Quote from: Martinus on July 28, 2015, 08:51:03 AM
Incidentally, the ethics of a machine making a decision about human life is not limited to automated weapons. Automated cars are another example - let's say you are driving in an automated car and you get two (non automated) cars coming from the other side. One is full of kids, who have a high chance of dying if you hit it, but you also get some chance of surviving; the other is a huge truck which will surely kill you. There is no way to avoid hitting either. Are we happy with your automated car making a decision who to kill (including, yourself)?
Are people working on self-driving cars that will scan nearby vehicles to see how many human lifeforms are in them?
Robots won't harm humans, or by inaction allow humans to be harmed.
Fuck robots.
We tried that in Europe from 1939-45.
Quote from: Norgy on July 28, 2015, 03:27:47 PM
Fuck robots.
We tried that in Europe from 1939-45.
:blink:
You were more like robot-enabling scum of the Earth neutrals. :hug:
I resent the notion that I am a descendant of robots.
My ancestors were puppets of the finest quality, thankyouverymuch.
Quote from: Martinus on July 28, 2015, 08:51:03 AM
Incidentally, the ethics of a machine making a decision about human life is not limited to automated weapons. Automated cars are another example - let's say you are driving in an automated car and you get two (non automated) cars coming from the other side. One is full of kids, who have a high chance of dying if you hit it, but you also get some chance of surviving; the other is a huge truck which will surely kill you. There is no way to avoid hitting either. Are we happy with your automated car making a decision who to kill (including, yourself)?
IF children_in_oncoming_car > 0 THEN swerve(right) ELSE swerve(left) ?
That's ridiculous. No one will ever program any line of code that makes a moral decision on whether humans should be killed or not. Cars are (even now) programmed to do everything to limit impact speeds. That's objective as it is just about Newtonian mechanics and there is no moral component to it. And computers can certainly make that decision much better than any human ever could.
Quote from: garbon on July 28, 2015, 02:32:57 PM
Quote from: Martinus on July 28, 2015, 08:51:03 AM
Incidentally, the ethics of a machine making a decision about human life is not limited to automated weapons. Automated cars are another example - let's say you are driving in an automated car and you get two (non automated) cars coming from the other side. One is full of kids, who have a high chance of dying if you hit it, but you also get some chance of surviving; the other is a huge truck which will surely kill you. There is no way to avoid hitting either. Are we happy with your automated car making a decision who to kill (including, yourself)?
Are people working on self-driving cars that will scan nearby vehicles to see how many human lifeforms are in them?
I am not sure - I sure hope so. If we are to let self-driving cars drive, they better be able to tell a difference between crashing into a bunch of bushes and into a bunch of pedestrians. :huh:
Which is, of course, rather far off from your hypothetical and the intention of my question.
Quote from: Caliga on July 28, 2015, 10:13:49 AM
I totally agree with them. I can't think of any reason we'd want to create a true AI anyway unless we do in fact want to go extinct. Human minds are pretty awesome.
They're weak and should be retired as soon as better materials become available.
Quote from: Caliga on July 28, 2015, 10:13:49 AM
I totally agree with them. I can't think of any reason we'd want to create a true AI anyway unless we do in fact want to go extinct. Human minds are pretty awesome.
I can think of a dozen reasons a fully automated economy would be way, way, way better than what we have now.
This is ridiculous. Robots will eventually be better at war than men, because robots have no feelings, they will obey the Rules of Engagement to the letter, they don't get angry, or fustrated, or tired of dealing with double crossing moonslims, or have personal problems back home that could influence their behavior and performance. Not to mention all the lives they will save by removing Soldiers from the battlefield.
Only retards that don't understand how flawed we humans are (not that I am human) expose these stupid and uneducated opinions.
I'm starting to think this Stephen Hawkins dude is not that intelligent after all. He cannot even see the emergence of the Humanai, the merge between humans and AI, the next step in human evolution.
Elon musk at least have Tesla Energy, with its new and incredible high performance batteries.
Quote from: Siege on July 29, 2015, 11:00:36 AM
This is ridiculous. Robots will eventually be better at war than men, because robots have no feelings, they will obey the Rules of Engagement to the letter, they don't get angry, or fustrated, or tired of dealing with double crossing moonslims, or have personal problems back home that could influence their behavior and performance. Not to mention all the lives they will save by removing Soldiers from the battlefield.
Only retards that don't understand how flawed we humans are (not that I am human) expose these stupid and uneducated opinions.
The man has a point.
QuoteI'm starting to think this Stephen Hawkins dude is not that intelligent after all.
And then he's back off the rails again :)
Situation: normal.
Robots that fight are one thing. What worries these luminaries(as well as myself) is robots that think.
If only we could get humans that think.
Quote from: Queequeg on July 29, 2015, 10:49:25 AM
I can think of a dozen reasons a fully automated economy would be way, way, way better than what we have now.
Let's hear them.
Rich people would no longer need to rely on poors for the slightest thing.
Quote from: Admiral Yi on July 29, 2015, 05:48:28 PM
Quote from: Queequeg on July 29, 2015, 10:49:25 AM
I can think of a dozen reasons a fully automated economy would be way, way, way better than what we have now.
Let's hear them.
Unlimited material production and enormous amounts of leisure time.
Quote from: Ideologue on July 29, 2015, 06:32:22 PM
Rich people would no longer need to rely on poors for the slightest thing.
Not revolting/voting to confiscate them.
Quote from: The Minsky Moment on July 29, 2015, 06:33:35 PM
Unlimited material production and enormous amounts of leisure time.
Unless one machine can make infinite machines with finite raw materials, I don't see how that is true.
Quote from: garbon on July 29, 2015, 07:25:05 AM
Which is, of course, rather far off from your hypothetical and the intention of my question.
I've seen Mart's hypothetical in several articles on the subjecty, it's definitely something that people in charge of writing the code for the cars, and those writing the laws for such cars, are thinking about.
Not this discussion again.
Our long term development leads to a post scarcity civilization.
Quote from: Peter Wiggin on July 29, 2015, 01:14:47 PM
Robots that fight are one thing. What worries these luminaries(as well as myself) is robots that think.
Robots that think? Who wants to build such a thing?
All people I follow in the robotics industry aim at highly specialized robots with narrow AIs.
People working on general AIs, like Google's Deep Mind project, expect to have the AIs in computers interfacing through virtual worlds. They see a strong AI on a robot as a waste of its capabilites.
Did you see the recent DARPA robot challenge?
They cannot even walk on an uneven surface...
Quote from: Siege on July 29, 2015, 08:47:59 PM
Did you see the recent DARPA robot challenge?
They cannot even walk on an uneven surface...
Just like Hawking. :ph34r:
Quote from: Ideologue on July 29, 2015, 06:32:22 PM
Rich people would no longer need to rely on poors for the slightest thing.
I am not too worried. Rich people still need status goods. We poor people can become the rich's status goods. See that rich man over there? He has so many underlings following him around, he must be rich.
Quote from: Siege on July 29, 2015, 08:38:52 PM
Not this discussion again.
Our long term development leads to a post scarcity civilization.
Post-scarcity civlization only makes sense in socialism. In capitalism, it will just mean absolute accumulation of capital in the hands of the few and/or revolution - that's worse than what we have now.
Quote from: Admiral Yi on July 29, 2015, 06:56:10 PM
Quote from: The Minsky Moment on July 29, 2015, 06:33:35 PM
Unlimited material production and enormous amounts of leisure time.
Unless one machine can make infinite machines with finite raw materials, I don't see how that is true.
Totally automated production implies the ability the produce more producers. Thus the only limit is raw materials. But that historically that has not been a binding constraint since the dawn of the industrial era. Since the evidence suggests that high levels of material affluence tends to be accompanied by low or even negative population growth, that probably would not change either.
Quote from: The Minsky Moment on July 30, 2015, 01:33:23 AM
Quote from: Admiral Yi on July 29, 2015, 06:56:10 PM
Quote from: The Minsky Moment on July 29, 2015, 06:33:35 PM
Unlimited material production and enormous amounts of leisure time.
Unless one machine can make infinite machines with finite raw materials, I don't see how that is true.
Totally automated production implies the ability the produce more producers. Thus the only limit is raw materials. But that historically that has not been a binding constraint since the dawn of the industrial era. Since the evidence suggests that high levels of material affluence tends to be accompanied by low or even negative population growth, that probably would not change either.
Woudn't we need space to put the producers in? So unless we keep miniaturising and/or colonising new planets (or we start culling the poor), such economy would collapse under itself in a Malthusian catastrophe in a few generations.
Planets are not neccessary, O'neil Cylinders are good enough.
https://upload.wikimedia.org/wikipedia/commons/9/94/Spacecolony3edit.jpeg
Quote from: Martinus on July 30, 2015, 01:31:58 AM
Quote from: Siege on July 29, 2015, 08:38:52 PM
Not this discussion again.
Our long term development leads to a post scarcity civilization.
Post-scarcity civlization only makes sense in socialism. In capitalism, it will just mean absolute accumulation of capital in the hands of the few and/or revolution - that's worse than what we have now.
Well from what the modern politics tell me socialism is just capitalism with slightly different numbers in a spreadsheet so we should be ok.
Quote from: Martinus on July 30, 2015, 01:31:58 AM
Quote from: Siege on July 29, 2015, 08:38:52 PM
Not this discussion again.
Our long term development leads to a post scarcity civilization.
Post-scarcity civlization only makes sense in socialism. In capitalism, it will just mean absolute accumulation of capital in the hands of the few and/or revolution - that's worse than what we have now.
What are you talking about?
Every successful nation on earth have build their economy while using capitalism, and only after being successful have become a welfare state socialist cesspool extremely slowing down their economic growth.
There is a reason why the Soviet lost.
Since we added welfare in the 1930s our economic growth has been extremely slow and anemic. :(
Quote from: Martinus on July 30, 2015, 01:39:06 AM
Woudn't we need space to put the producers in? So unless we keep miniaturising and/or colonising new planets (or we start culling the poor), such economy would collapse under itself in a Malthusian catastrophe in a few generations.
There's an awful lot of land - it would take a inconceivably vast expansion of production for space constraints to bite hard.
So yes literally speaking production isn't "unlimited" - practically it is though.
Quote from: Siege on July 30, 2015, 08:45:01 AM
What are you talking about?
Every successful nation on earth have build their economy while using capitalism, and only after being successful have become a welfare state socialist cesspool extremely slowing down their economic growth.
There is a reason why the Soviet lost.
The question is what role is left for a capitalist to perform in a purely automated economy. Assuming the automated agents are capable of developing improvements on their own, and assuming automated agents are capable of monitoring demand and allocating resources (including capital) - it's not clear that there is any.
Quote from: The Minsky Moment on July 29, 2015, 06:33:35 PM
enormous amounts of leisure time.
This is actually a horrible thing, not a good one. People who have too much leisure time often spend it thinking, and we don't want people thinking too much because then they come up with crazy ideas, like black people should be killed in their churches or random people should be killed in movie theaters, or children should be killed in their elementary school, or Americans should be killed in the World Trade Center.
The only crazy ideas involve murder? :yeahright:
Quote from: The Minsky Moment on July 30, 2015, 12:15:04 PM
Quote from: Siege on July 30, 2015, 08:45:01 AM
What are you talking about?
Every successful nation on earth have build their economy while using capitalism, and only after being successful have become a welfare state socialist cesspool extremely slowing down their economic growth.
There is a reason why the Soviet lost.
The question is what role is left for a capitalist to perform in a purely automated economy. Assuming the automated agents are capable of developing improvements on their own, and assuming automated agents are capable of monitoring demand and allocating resources (including capital) - it's not clear that there is any.
Misnky, serious answer:
I agree the economic system of the future might be different from capitalism, however I don't see how we can get to supply-side future of abundance without having free market capitalism growing the economy today, creating the opportunities for inventors and investors to develop and commercialize their products.
Socialism and the welfare state just don't have the economic growth to propell us into an abundance 360 society, let alone to the post scarcity utopia.
The things that technology promises, the digitalization, dematerialization, demonetalization, and democratization of the economy, are only possible with the evolutionary selection of products and services that only a free market economy can make possible.
Any other economic model selects products and services based on political decisions rather than consumer selection.
I'm sorry. I don't think I am expressing my point coherently. I hope you understand the idea I am trying to convey.
Yes the answer is coherent. Probably wrong but coherent. ;)
Actually there is a decent argument that a good social safety net encourages entrepreneurialism by limiting the downside of failure.
Quote from: Admiral Yi on July 29, 2015, 06:56:10 PM
Quote from: The Minsky Moment on July 29, 2015, 06:33:35 PM
Unlimited material production and enormous amounts of leisure time.
Unless one machine can make infinite machines with finite raw materials, I don't see how that is true.
Unless you never recycle, I don't see how this isn't true. One doesn't need raw materials to produce, one only needs feedstock, which could be raw or intermediate materials.
Anyway, on the topic, most of the LARS people are talking about are really just advancements on the mine; they are triggered by a sensor, and then they act. They are no more "summary execution" than a machine gun blind-firing at potential enemy positions.
What AI does for us is allow us to avoid the land mine/blind fire/potential-killer-robot approach by making the killer robot purely reactive; unless fired upon, it won't fire. But if it is fired upon, it can response with a speed and accuracy that, though it precludes the ability of humans to authorize return fire, also doesn't involve noncombatants. The actual combatants firing on it couldn't move quickly enough to escape return fire.
Imagine that, in Mogadishu in 1993, the US deployed a battalion of reactive killer bots instead of the Delta and ranger guys. Aidid's forces could engage and die or not engage and allow Aidid's lieutenants to be captured. In either case, the mission is accomplished because the mission wasn't intended to kill a thousand militia.
Now, these things would have to be proof against reprogramming (Brazen makes that excellent point in her article) and used with care, but casualties wouldn't be a major issue and the benefits seem to me to greatly outweigh the risks.
Quote from: Siege on July 30, 2015, 08:45:01 AM
Quote from: Martinus on July 30, 2015, 01:31:58 AM
Quote from: Siege on July 29, 2015, 08:38:52 PM
Not this discussion again.
Our long term development leads to a post scarcity civilization.
Post-scarcity civlization only makes sense in socialism. In capitalism, it will just mean absolute accumulation of capital in the hands of the few and/or revolution - that's worse than what we have now.
What are you talking about?
Every successful nation on earth have build their economy while using capitalism, and only after being successful have become a welfare state socialist cesspool extremely slowing down their economic growth.
There is a reason why the Soviet lost.
This is simply not true. Capitalism is a modern phenomenon. It was theorized by people beginning in the 18th century and was really put into place in Europe and American in the 19th century. Vast empires rose declined and collapsed longer before capitalism. We can look at some countries and retrospectively say they had elements of capitalism, but I don't think we can say the actually had capitalist economies.