Slashdot Mirror
Answering Elon Musk On the Dangers of Artificial Intelligence
Lasrick points out a rebuttal by Stanford's Edward Moore Geist of claims that have led the recent panic over superintelligent machines. From the linked piece: Superintelligence is propounding a solution that will not work to a problem that probably does not exist, but Bostrom and Musk are right that now is the time to take the ethical and policy implications of artificial intelligence seriously. The extraordinary claim that machines can become so intelligent as to gain demonic powers requires extraordinary evidence, particularly since artificial intelligence (AI) researchers have struggled to create machines that show much evidence of intelligence at all.
But they're not actual AI. I mean, you might as well outlaw cruise missiles or why not claymores and mines?
A drone killer doesn't just kill anything in its zone. It has a threat profile its looking for and so far that profile has been so specific that the actual literal target is specified. aka... THAT truck or THAT house or whatever. Its not "stuff that looks like a truck" or "stuff that looks like a house" or "people".
its specific to a DUDE.
now the sort of stuff the military is talking about automating are things like drone tanks that will deploy to zones and then shoot anything that moves in that area... and potentially be able to tell the difference between stuff so they don't just shoot anything. But the problem with not shooting anything that moves is that it would be very easy for a human being to walk up to one of those drone tanks with a big sachel anti tank bomb, put it on the tank and walk away... Boom... end of tank. That's likely not ideal. So you start running into concepts like drone tanks as area denial weapons like mine fields. So they go to an area and they will totally kill anything in that zone that does not squak an IFF. So you can use them for base defense, holding an enemy in pinned by putting the drone tanks on one side of their base and then moving in manned forces on the other side to pincer them between the two. The drone tanks being used to cut off escape.
There are ways to use this...
Another play on the same concept is air defense attack drones. They would engage ANY airplane that enters the zone. Ideally not a commercial airliner full of orphans. But the idea is the air defense drones can operate the way our other drones operate. Long endurance. hanging out over a target all day... Patient. The ground attack air drones are probably going to remain human targeted. I don't really know why we'd change that. The human operations use the close support bombers as artillery in the sky basically. Guy on the ground says "kill that"... and the drone does. And then of course there is the naval model... drone torpedoes that just hang out and wait. Ship enters the waters without squawking an IFF... BOOOM. Again, area denial. Don't want to get hit... stay out of those waters. This could be used to defend harbors in contested territory, seeded behind enemy lines to sow chaos as shipping gets hit with random torpedeos from subs that aren't there. Also very dangerous engagements against well defended naval targets. Send the drone torpedo in slow, deep, and silent. Moving a couple miles an hour... just edging in there... and then when it detects the target... slowly slowly... BOOM. or possibly it latches on to the side of the ship and does something else. The damn thing could hack the enemy ship's network for all I know. Whatever is deemed desirable.
This notion of the terminator killer robots is not how they'll be used. We don't trust them and they're not that smart.
I've decided to stop wasting my time responding to AC trolls/sockpuppets... so if you want a response from me... login.
The first problem when arguing about the dangers or chances of AI is agreeing on what AI is even supposed to be. Laymen will most likely be referring to "strong AI", meaning, AI with human capabilities, such as creativity or even consciousness, whereas professionals will probably think of AI in more practical terms, as in a software that can solve a set of limited or very specific problems by making informed, "intelligent" decisions.
Today and in the foreseeable future, we will only get the latter, weak AI. People panicking about the dangers of AI usually have strong AI in mind. Professionals don't take them seriously because they know that strong AI is not even on the horizon.
Problem is that there are numerous ways even weak AI can go very, very badly. There was the big stockmarket crash some years ago, caused by automated trading algorithms. Think self-driving cars that have been hacked or have faulty programming. Think automated defense systems that get fed wrong data or malfunction.
These are the kinds of AI issues to worry about. The Asimov-style superhuman intelligence taking over is not something to be concerned about at the moment.
Presently, The USA has no landmines deployed anywhere in the world.
Except in Vietnam, Korea, all the fuck over Europe, Afghanistan, Iraq and Iran.
Just because they left them behind, doesn't mean they're not 'deployed'.
One has to wonder. How would the public react if, say, the Mexican government used a drone to kill a global criminal in Los Angeles. Even better, what if they also took out 2 innocent fathers, 1 mother, and 3 kids while killing the bad guy?
I'm going out on a limb here, but I'll bet the American public would react a whole lot differently than they do when an American drone takes out 1 maybe-terrorist + a wedding party in Pakistan.
With a non-autonomous weapon, the person who pulls the trigger is basically responsible. If you're strolling in the park with your wife, and some guy shoots her, well, he's criminally liable. If some random autonomous robot gets hit by a cosmic ray and shoots your wife, nobody's responsible.
This is a huge issue for our society, because the rule of law and criminal deterrence is based on personal responsibility. Machines aren't persons. The death penalty for a machine is stupid (watch out, robot, if you kill someone we'll take out your batteries!). The number of ways that things can go wrong without the owner of the machine having a reasonable amount of liability is huge.
What if the autonomous weapon malfunctions in the field? Is the owner responsible for having deployed in that particular location? Is the manufacturer responsible for the bugs that occur? What if the machine is operating outside of recommended parameters? What if the machine was hacked, and the bug occurs due to a faulty communication issue, ie the message was sent to authorize targeting your wife, but then a fraction of a second later another message was sent rescinding the order, but the message was garbled or never arrived due to a netwoking delay in transit on Amazon's cloud servers? What if the machine's owner deploys thousands of vermin killing robots around the city without incident every day, but it just happened to kill your wife because she was misidentified as a rodent?
The fact is that AIs and autonomous robots have no legally useful place in society (unlike nonautonomous robots). There is almost no deterrence value in threatening an owner with fines (how much is reasonable in the rodent example?) and there is no value in destroying the offending machine (an autonomous machine is not alive, and it may be the identical model from a manufactured run of 1 million products, so what's the point of scrapping that one unit?). There is no point is blaming a random customer who bought the machine and probably has no clue at all how it operates or how to detect malfunctions. And you can bet that the manufacturing chain is full of lieability disclaimers and insurance companies will pass the buck. So what hope is there for avenging your wife? And if it goes to trial (against whom?) how long and how much cost will be spent for an uncertain outcome?
The ethical issues surrounding blame are serious, and at the risk of going slightly off topic, they are similar to the issues of terrorism. If a suicide bomber blows himself up in a crowded place, you can't pick up his pieces and stick them in jail. Nothing you can do to him has any deterrent effect, and going after his family or friends is, at best, a legal nightmare and an ethical problem. The issues surrounding autonomous machines are a bit like that, because, well, the fact that it's an *autonomous* machine means that no human being was actually pulling the trigger or directly making the choice to shoot.
Out of all of the weapon-specific hysteria (and there has been a lot of it--white phosphorus, thermobaric bombs, depleted uranium, etc.), the anti-landmine one might be the most dangerous.
Obviously, they do have a good point, what with the disasters in Indochina and elsewhere. However, those were cases of non-self destructing anti-personnel landmines placed in third world nations. The situation is / would be quite a bit different with anti-tank mines, self-deactivating or remote-deactivating mines, and/or mines placed in developed nations that have the resources to keep people out and clear the minefields later on as needed.
Why is this all worth mentioning? One word: Ukraine. In a situation where one side in a conflict desperately wants to fortify their defenses but doesn't want to risk alarming the other side (or giving them a plausible pretext to feign alarm), landmines are one of the few stationary weapons available that can thwart or at least seriously slow down an invasion. Instead of all this deeply worrying Cold War-type bravado of military exercises and NATO rapid response plans in Eastern Europe, just mine the fuck out of their borders. Putin could act huffy and offended if he wants, but people will realize it is a clearly not an aggressive action.
The people raising the "alarm" are industrialists who want to divert attention away from the *real* impact of the current trends in automation - the replacement of human workers by robots. I'm really tired of people talking about super intelligent AIs who for some reason resemble us only in an irrational desire to destroy things when the real issue is how are we going to re-structure our society when 50% of the population doesn't have jobs? Just look at the countries where the unemployment rate goes north of 10% or 15% - all manner of criminal behavior ensues as people scramble to find a way to put food on the table. What do you think happens when it hits 40% or 50%? Unless we put some serious thought into how our society is going to provide an acceptable standard of life for 50% of its jobless population, it is going to get ugly really fast.