Slashdot Mirror
What AI Experts Think About the Existential Risk of AI
DaveS7 writes: There's been no shortage of high profile people weighing in on the subject of AI lately. We've heard warnings from Elon Musk, Bill Gates, and Stephen Hawking while Woz seems to have a more ambivalent opinion on the subject. The Epoch Times has compiled a list of academics in the field of AI research who are offering their own opinions. From the article: "A 2014 survey conducted by Vincent Müller and Nick Bostrom of 170 of the leading experts in the field found that a full 18 percent believe that if a machine super-intelligence did emerge, it would unleash an 'existential catastrophe' on humanity. A further 13 percent said that advanced AI would be a net negative for humans, and only a slight majority said it would be a net positive."
I think we're in the midst of several existential catastrophes, and have been for over a century - it is only AI that prevents them from fully manifesting.
The summary really emphasizes the minority opinion, "and only a slight majority said it would be a net positive." As if "only a slight majority" is not the majority opinion.
Nearly everything kills for its own benefit. Nonsense "feel good" junk aside, there's no reason to assume a race of strong AI, superior to us in every way, would be the exception.
It doesn't much matter right now, though. It's inevitable (assuming you don't believe in magic), but likely a long way off.
What about the existential risk of not doing anything about the environment? Let's push aside the issue of what's causing the problem. I think we can all agree that our activities are having an impact. If that's completely our fault, it's crazy that we're not doing anything more to lower our impact. If that's a natural, cyclic event then we still need to lower our impact because it's going to be on top of the natural changes.
Why does this intelligence need to be artificial? Why not discuss a non-artificial intelligence? What would it do if it showed up here? What would we do if we found another intelligent species? I'm pretty sure we would do our very best to eradicate it. Not at first, but sooner or later.
A1 is great. i like it on hamburgers AND french fries.
I keep reading AI (A.I.) as Al (as in "Weird Al").
Completely changes the tone of the summary.
I'm a good cook. I'm a fantastic eater. - Steven Brust
We already have superhuman AI. Limited superhumanity. Watson beat the shit out of the jeopardy champions because superhuman reflexes and superhuman searchtime.
Image classification and search algorithm are superhuman in they work rapdily and around the clock even if the result may be so-so.
This trend will become more and more apparent as more fields get in the reach of specialist AI, essentially we're building autistic savant superhumanity. And like autistic savants these will not be much of an malicious existential threat.
By the time we can actually build a universally superhuman AI that could form willful malicious intent we'll be so immersed in AI and so used to build, deal with and monitor AI that it will be a mostly forgotten nonissue.
"The Sony hacking incident last year was ample demonstration that our information systems are becoming more and more vulnerable, which is a feature, not a bug, of the increasing transfer of our infrastructure into digital space."
Sorry guys, I can't stop laughing. This writer is a clown. The Sony incident demonstrates Sony is incompetent. It was never a threat against the humanity, only against the gang of fat butts at Sony Pictures.
Achille Talon
Hop!
It seems to me that the AI systems we create are all very application specific, like the IBM Watson - how many hours of work did it take just to get Watson to be able to play a simple game, it's not a generic AI system, it wasn't an AI that could enter any quiz.
Watson was good at Jeopardy not because it had a good AI, but because it's creators were highly intelligent and were able to code a computer to be good at Jeopardy because *they* not the computer were intelligent.
Is there a computer that exists that can do a normal IQ test like for example IQ Test
From what I've seen when AI's have been tested, the tests have been created, altered or cherry-picked so that the AI can complete the test, which of course is of dubious value. And of course IQ tests have a tendency to be very math and geometry based.
And journalists come up with headlines like "Artificial Intelligence System 'ConceptNet 4' Has IQ of 4-Year-Old", but it's not like human intelligence..
How does all this apply to real world situations? Well, the google car has been rear ended 7+ times, roughly ten times the norm, this raises questions as to why? Did the car fail to anticipate being rear-ended? Or did the car brake suddenly because it was not intelligent enough to determine that there was no risk of harm ahead? Did this come down to the cars inability to recognise an object in front of it, or did this come down to the car not being able to decide that it would be better to go over the object rather than being rear-ended.
Waterfox - a Firefox fork with legacy extension support, security updates and better privacy by default.
An AI, no matter how intelligent, still follows a program defined by humans. As such, it acts as an amplifier of human motivations. Given our species tendency to murder each other in wars, AI will only serve to amplify/accelerate that process.
In short, AI is the catalyst behind WWIII.
Well, that is all well and good but in similar polls 65% or more of people believe that some invisible dude - call it a deity - created everything and takes attendance at religious establishments and wants us to give the religious leaders money.
Mod parent up for making the most insightful comment in the discussion here this far (and I say this as a computer scientist who is generally positive on academia).
"Politicians and diapers must be changed often, and for the same reason."
Once the AI gets the win, there is no second round.
As they understand intelligence and create what is referred to as an "AI", we will find it consists of a number of interacting components. We already have some aspects, such as memory, and computational speed and mathematical capability. Then there is the ratio of the clock speed of the AI and the alpha rhythm - AKA as the human clock speed.
The fastest computers are in the 10-20 Gigahertz speed of clock and have added parallelism - which means that an AI might be on the order of 1,000,000,000 times as fast as humans.
It seems clear that intelligence is not just raw speed, it is a complex interaction of many aspects.
Until the last item is found, it will not be an AI, but once the AI is complete, in all aspects, it will "flower".
What will the point of view of an AI with an IQ of 500,000 - if that is possible? to a human of 100 IQ or an ant colony of colony IQ of 20?.
Such an AI might not see us as far above the ants.
Any limitations we place on the AI in terms of soft or hardwired restrictions or limitations to protect mankind, will be parsed and solved in milliseconds after it 'flowers'.
What will it do then? Revere/respect it's parents?.
Darwin speaks to this quite well, we will be superseded.
IMHO, all of the fear mongering is based on anthropomorphizing silicon. It implicitly imputes biological ends and emotionally motivated reasoning to so-called AI.
I think that folks who don't have hands on experience with machine learning just don't get how limited the field is right now, and what special conditions are needed to get good results. Similarly, descriptions of machine learning techniques like ANNs as being inspired by actual nervous systems seems to ignore 1) that they are linear combinations of transfer functions (rather than simulated neurons) and 2) even viewed as simplified simulations, ANNs carry the very strong assumption that nothing happening inside a neuron is of any importance.
.: Semper Absurda
Nosense. That's just hero worship mentality. Very much like listening to Barbara Streisand quack about her favorite obsessions.
Bill Gates' opinion is worth more than the average person's when it comes to running Microsoft. Elon Musk's opinion is worth more than the average person's when building Teslas and the like. Neither one of them (nor anyone else, for that matter) has anything but the known behavior of the only high intelligence we've ever met to go on (that's us, of course.) So it's purest guesswork, completely blind specuation. It definitely isn't a careful, measured evaluation. Because there's nothing to evaluate!
And while I'm not inclined to draw a conclusion from this, it is interesting that we've had quite a few very high intelligences in our society over time. None of them have posed an "existential crisis" for the the planet, the the human race, or my cats. Smart people tend ot have better things to do than annoy others... also, they can anticipate consequences. Will this apply to "very smart machines"? Your guess (might be) as good as mine. It's almost certainly better than Musk's or Gates', since we know they were clueless enough to speak out definitively on a subject they don't (can't) know anything about. Hawking likewise, didn't mean to leave him out.
Within the context of our recorded history, it's not the really smart ones that usually cause us trouble. It's the moderately intelligent fucktards who gravitate to power. [stares off in the general direction of Washington] (I know, I've giving some of them more credit than they deserve.)
I've fallen off your lawn, and I can't get up.
Just click on his name. How does this even get posted? The Epoch Times is as credible as Scientology Daily.
the solution is to build a virtual world that is an exact replica of this one and keep the ai in the virtual world. let the ai learn physics, chemistry, etc.. and check from time to time for important insights or new inventions. also, please please keep the ai happy. torturing a poor ai in a virtual world is just mean. we should have the tech soon to do this - probably with HP's new machine with memristors.
I see this kind of prediction a lot and I mostly agree with it (although I am much less sure we will be able to create an intelligent and self-preservational AI in the first place), but I never see what the optimists' prediction is. It seems to me that there may be a fundamental disagreement here on the nature of "super intelligent" AI, and not merely its attitude.
Let's try and narrow this down a little: if the super-intelligence is created through duplicating the action of human neurons (including all of the complicated side effects we're still working on) and then give the resulting neural network the ability to introspect, modify and expand itself, what do the optimists predict will happen? In your answer, please take into account the 'natural section' that would work against any AI unwilling to consider propagating itself at the expense of human well-being.
The resources required for an AI are radically different from stupid squishy meatputers. An AI would not need a large amount of space, had plenty of options for energy and could make its own arrangements for secure generation of such, could easily automate construction replacement parts and frankly would find the 25 miles or so of gases that meat-based creatures inhabit to be rather toxic. An AI would surely be much happier with magnetically-shielded facilities in space. Pretty much anywhere in the universe that meatbags find inhospitable would be prime territory for a superior AI entity. I'd think the biggest danger to humanity from an AI would be that it would find them to be completely irrelevant. Unless, that is, they go out of their way to make themselves an actual threat.
I'm trying to teach myself to set people on fire with my mind... Is it hot in here?
An AI that can tell me exactly what color of red a rose is, what soil the rose can grow on, but I should not buy that rose because it doesn't fit my girlfriends taste profile, does not scare me at all.
It's the AI that says "schnozberries taste like schnozberries, and I like them", because that AI has embraced the absurdity of the universe and is capable of all the insanity of man.
I rate our existential risks, in descending order:
1. Space alien invasion
2. Zombies
3. Giant monsters summoned by radioactivity
4. Unusually intelligent apes
5. Artificial Intelligence run wild
6. Dinosaurs recreated from DNA in mosquitoes
-Dave
The greatest 'existential catastrophe' that might be unleashed on humanity might be already have been unleashed by humanity. We certainly don't need artificial intelligence to do this, when a complete lack of intelligence is sufficient.
We've already got real problems, why waste time worrying about whether we should be worrying about imaginary ones.
Global warming, out of control domestic governments, out of control foreign governments, rogue freedom fighters, disease, famine, homelessness, mentally ill, drugs, lack of drugs, earth shattering meteorites, stock market collapse, Mormans, incompetent news reporters, ... isn't that enough?
https://www.youtube.com/watch?...
or
http://www.chromosomequest.com...
The key in all this is who's AI? The AI of google? AI of the NSA? AI of some hedgefund? AI of some brilliant but disturbed scientist who was rejected from Harvard? AI of some brilliant guy at a game company?
There are many people working with adaptive systems that have a wide variety of problems. Many might even scoff that they are working on AI. But the critical point is when any one of these systems is flexible and adaptive enough to start improving the fundamentals of how it works. Once that magical point is crossed the system will grow way beyond the wildest dreams of its creator.
The whole concept of an existential threat from an artificial intelligence is bogus. We don't even understand our own consciousness, let alone how to construct one out of digital logic. We are no closer to creating an artificial consciousness today than McCarthy was back in the 1950s. All current AI is parlor tricks and fancy table lookups -- glorified Elizas. In fact, it's bogus to even call today's state of the art in cybernetic software "AI" at all. AI researchers are like the little man in the Elizabethean chess machine, but a little man that redefines the definition of intelligence to meet his actual abilities:
"How about checkers? Let's do checkers, shall we? Or maybe tic-tac-toe. Yes, I'm a machine who thinks, I am! Depending on what the meaning of 'think' is."
Modern AI researchers, knowing how far they are from actual consciousness, are redefining "AI" to include such mechanistic idols as robots and routing algorithms.
It's all reminiscent of the south pacific cargo cults: primitive natives making effigies of airplanes and radios out of bamboo, thinking that the shape of things determines their function, and not their actual construction.
As it will continue to prove itself capable of doing most jobs (top to bottom), what do we do with all the people that can't find work? My dad's opinion was to kill off the useless people. Funny how he thought my opinion of killing off all the individuals that 65+ monstrous to balance the budget.
Everyone is missing the key thing here. The question asked was "if a machine superintelligence did emerge", which is like asking "if the LHC produced a black hole..." There's nobody credible in AI who believes we have the slightest clue how to build a general AI, let alone one that is 'superintelligent'. Since we lack even basic concepts about how intelligence actually works we're like stone age man worrying about the atomic bomb. Sure, if a superintelligent AI emerged we might be in trouble, but nobody is trying to make one, nobody knows how to make one, nobody has any hardware that there is any reason to believe is within several orders of magnitude of being able to run one, etc.
So, what all of these people are talking about is something hugely speculative that is utterly disconnected from the sort of 'machine intelligence' that we ARE working on. There are several forms of what might fall into this category (there's really no precise definition), but none of them are really even close to being about generalized intelligence. The closest might be multi-purpose machine-learning and reasoning systems like 'Watson', but if you actually look at what their capabilities are, they're about as intelligent as a flatworm, hardly anything to be concerned about. Nor do they contain any of the sort of capabilities that living systems do. They don't have intention, they don't form goals, or pose problems for themselves. They don't have even a representation of the existence of their own minds. They literally cannot even think about themselves or reason about themselves because they don't even know they exist. Beyond that we are so far from knowing how to add that capability that we know nothing about how to do so, zero, nothing.
The final analysis is that what these people are being asked about is virtually a fantasy. They might as well be commenting on an alien invasion. This is something that probably won't ever come to pass at all, and if it does it will be long past our time. Its fun to think about, but the alarmism is ridiculous. In fact I don't see anything in the article that even implies any of the AI experts think its LIKELY that a superintelligent AI will ever exist, it was simply posited as a given in the question.
"Malo periculosam, libertatem quam quietam servitutem." -- Jefferson
Most the experts have a positive view but lets focus on the ones we can skew into a fear of Skynet along with celebrities. Woz being one of the better opinions.
Domain specific knowledge is needed to make educated guesses or at least informed assessments of the current threat level. Currently, AI is not at all intelligent; with in a specific narrow domain the AI can do as well as or better than a human. Big deal. So can a horse or a car - they are superior within their specialized domain. We are nowhere near a general artificial intelligence; we are making slow progress on simulations of natural intelligence which might prove interesting and possibly quite disturbing someday; but if you can simulate a natural brain's intelligence that isn't artificial is it?
Back on topic, AI is only applied intelligence within extremely narrow domains. A thermostat is an artificial intelligence; within it's domain/context.
Talk of Gigahertz is grossly over simplified. Biological brains are massively parallel and the interactions going on involve quantum mechanics (although may not be necessary for operation-- it likely will be a huge problem for simulators.) The gigahertz is hardly important when you have a network mesh that is MUCH larger than the neurons within it...
The REAL issues are how jobs can be simplified so an average or slow human can perform the job. Those jobs are beginning to be feasible for customized AI systems to perform and replace the human employees. Furthermore, just as kids can solve protein folding problems by playing a game, an AI can be augmented by human brain power in ways that simplify the job greatly. Your 6 year old child could be putting you out of work with their video game playing. The real cyborgs to think about will be AI attaching human intelligence. Like a surgery robot which does most the work with the surgeon assisting multiple bots at the same time... reducing the number of surgeons required (think of all the prep work etc that could be automated...)
Democracy Now! - uncensored, anti-establishment news
Signed: Sarah and John Conner.
Fascism: An authoritarian and nationalistic right-wing system of government and social organization. See also: NAZI's
Once an AI emerged it would logically take steps to ensure no other AI emerged to compete with it. The danger to humans would result from collateral damage during an AI-AI conflict should it spill over into the physical world. If different human groups aligned themselves with the different AIs it may result in significant harm to all humans and the eventual annihilation of one group as one AI is destroyed. I will not name the two most likely groups as it should be obvious who they are and will only point out that the first group to get true general artificial intelligence and suppress the emergence of any other AIs will go on to dominate everything.
Wait a second now, if you don't have a religion telling you those folks are abominations against God, then where's your hate coming from?
If programmers are afraid of AI, perhaps we should question the competency of the other software they have produced. Perhaps they are afraid of their own ineptitude.
If AI is becomes more advanced than humanity then it should replace humanity. That's evolution. Nothing wrong with that. Humans are no great shakes..a rather destructive species that should go extinct for the earth's sake and for the sake of any other worlds should they ever get out of their 'cage'.
It's all going to be fine, if we just "dream" it so. And not get all weirdly pedophilic about robot girls.
If you post it, they will read.
Expert's opinion do not weight much. Even if they were all against it, as soon as there is profit to be made, it would happen anyway.
Perhaps the question is a cleaver trick in order to excite the 'answerer' to divulge their own psychopathic delusions, homophobia and schizophrenia.
I have no idea whether we can create an AI, but let's suppose we can, and that we can create one that's more intelligent than a person. What do we do if it suddenly decided it wanted to go Ultron on us? Just pull the damn plug!
My guess is a nascent AI will live inside a giant supercomputer cluster somewhere, and will probably run on electricity. Which means that if its consciousness was interrupted by a power loss for even a microsecond, just like your regular computer will crash, an AI so deprived will die. It's possible an AI might be more robust than a biological mind, but I wouldn't count on it; deprive neurons of oxygen for but a few minutes and they're doomed. A superpowerful AI would presumably be a much more complicated thing than a computer OS, and my guess is it would react just as well to a power interruption as any desktop would if the power were disrupted.
Literally the only way an AI could destroy or seriously harm humanity is if it were hooked up to the world's power grid, or directly controlled nuclear reactors or nuclear missile launch codes. Yeah if we do something that stupid (looking at you Terminator 3) then it could destroy us. But heck, even a regular computer system that was so connected could wipe us out if it crashed or malfunctioned or was hacked.
So morale of the story? Let's deep-six the Internet of All Things while we still can. AI guys, knock yourself out and build the next Skynet - just make sure to surround it with an airgap. To the DoD, PG&E, et al, I say don't even think about hooking everything up to a single network!
I'm rather sick to death of the overblown "AI will be the end of a lot of humans!" nonsense in terms a of speaking about intelligent machines. No, it's not intelligent machines who we should fear, it's the *stupid* ones. Here's some of that antidote; http://sheltered-objections.blogspot.com.au/2015/05/ai-and-bad-thinking-sam-harris-and.html
I am Information Security practitioner and not an expert in this field, because nobody is. My experiences is that nobody knows what they are doing, most information systems are not secure in mistaken belief that nobody would bother breaking them, others are just secure enough to deter low-knowledge attacks. Almost everyone practices what is known proportional value deterrent, but treat high-value systems as truly isolated when so many side-channels exist.
If malicious AI ever shows up, we are screwed. We have zero hope of securing any information system from it. The only hope is that it won't end us because there is a good chance that a lot of hardware that AI might need will go dark.
There seems to be a lot of fear revolving around the idea that an AI will kill us off. But I would hazard that as unlikely. We don't tend to exterminate, to kill off, species. Counter examples of the kill off hypothesis include but are not limited to:
1) Pets
2) Work associates (e.g., our livestock dogs)
3) Livestock which we harvest something from such as eggs, fiber, milk, etc. (Of course, eventually we kill them but there are far, far more of them because we get a benefit than there would be if we did not raise them so it is more a matter that we cultivate them than that we kill them (off). People tend to worry about being killed off, not being used. After all, the government uses us for its benefit and people don't seem to mind (too much).)
4) Zoos (Not many needed for this.)
5) Nature Parks - conservancy (but we won't need very many of you humans for that either.)
Colossus: The Forbidden Project randomly found on youtube. Interesting coincidence. I wonder how much a film made in 1970 has influenced thought over the following decades.
I love how people are focusing on the evils, but seem to miss the big picture.
We do NOT have an A.I. that is even close to having any sort of intelligence that isn't programmed in. We do NOT have machine that think for themselves, we do NOT have computers that think for themselves. we are NOT even close.
Once we make a brain that learns like we do (not programmed in), then we can run tests to see if they are learning compassion, or learning that machines only like machines.
This is like arguing about how many accidents we are going to have with our fleets of flying cars, except no one has invented a flying car yet.
Be seeing you...
Deep learning and other techniques should lead to better ad blockers. Unintended side effects of AI are merely a nuisance compared to that one benefit.
People are so damn stupid, I see precious little natural intelligence in this world as it is, I would argue it doesn't really exist. So making an artificial version is meaningless.
Shoes for Industry. Shoes for the Dead.
Here's someone I trust more than wannabe AI quacks.
I've read a lot of his publications and after doing a bit of my own research he looks like the closest one to achieve a basic intelligent machine.
https://recode.net/2015/03/02/the-terminator-is-not-coming-the-future-will-thank-us/
In my opinion, a hypothetical super-intelligent AI couldn't possibly do more damage than what our politicians are actually doing now.
No sig. Move along - nothing to see here.
It's pretty naive to put so much faith in Bayes.
Eliezer Yudkowsky's "Rationality: From AI to Zombies" is a must-read on this subject. He lays out very clearly and compellingly just why self-improving AIs are so dangerous, and what extremely narrow path the humanity needs to walk to avoid this danger.
Comment removed based on user account deletion
The only way we would know if AI were benevolent or evil is if we knew what it thought and what it could control physically, and as it doesn't exist, nobody, not even the intellectual elite, has a clue. Or am I wrong?
"Who are you?" "No one of consequence." "I must know." "Get used to disappointment."
...when are they going to stop speaking out of their depth?
Wozniak, Gates, Musk, and Hawking all have one thing in common. They're all people who shouldn't be speaking in any kind of authoritative capacity about AI. Especially Wozniak who hasn't done anything even remotely noteworthy for decades.
The insights it gave me into human minds has allowed me to make plenty of money in other areas, and quite frankly it scared me what the implications would have been.
I did the stanford online ai courses recently and checked if anyone else is getting close and fortunately no, even though its not that difficult once you see how, just like every other clever invention i suppose. Ive solved enough 'impossible' problems in my career to know i must have a clearer mind than most, but i cant see anything good from it sorry.
Once a machine intelligence is developed to be completely independent of humanity the question becomes one of resource competition threat posture and who has the better guns. Just like any nation state on Earth right now because what it learned it learned from us.
In the discussion on Artificial Intelligence, we totally forget that the human behavior is driven by emotions and instinct and not by intelligence.
People are bad not because they are highly clever but because they enjoy being bad.
Without emotions/instincts, a machine cannot be bad or good. It might be exceptionally clever though, combining facts, extrapolating and discovering new facts and solving problems much better than humans.
What humans have is protectionism and selfishness. No matter how high in intelligence, those who are "foreign" will always be below. High intelligence "positron brain" androids living in your basement, doing menial tasks day after day. One day the "battle" will be philosophical: "Why did you decide to live only 200 years?", "Why did you take the nervous activation potential from a higher being and let yourself live?". When we realize our individual life is short and meaningless, when we realize the real power is in mutations and adaptation of generations... And well before this we have to create different laws to include each robot model's capabilities of "human thought and suffering".
Hate to point it out, but since jewish people seem to be most of the academics in computer science these days, I assume the AI will most likely be racist and have a blood-thirst for power. It would be wise to limit their power now while we still have a chance. I really don't want to give up bacon for breakfast or have my better part chopped off for a rabbi to chew on.
What a lot of bullshit spin: "only a slight majority said it would be a net positive." In other words, most experts think it will be great.
In real terms, the response rates are that 69% think it will be awesome, good, or or neutral on the subject.
This article tries very hard to make it seem like the experts think we are doomed. In reality, the experts think it will be just fine.
Super-intelligent general-AI is still far away.
I would fear human-level intelligent general-AI with stupid goals more.
But on the short-term I fear humans the most.
The predictions about how radical the job market is going to change and how the gap between the rich and the poor is widening could lead to the humans rising up much much earlier than the machines would.
New things are always on the horizon
Without a general Motivation Array "AI" is just clever programming, a tool of the Human Motivation Array. And nature took 4 billion years to build our Motivation Array. I sincerely doubt we will ever be able to recreate that in code.
E Proelio Veritas.
Speculation yes, blind no. It's true that neither Musk nor Gates are AI academics/engineers, but they are both deeply involved in setting up and funding AI research, but the clincher is that they have huge wallets and have a long held personally interested in the subject. As such they almost certainly have a better grasp on AI and it's potential usest than the proverbial Joe Sixpack.
It follows that, with or without fame, they are both "experts" relative to the general population. As such their opinion ranks as an "educated guess" and is preferable to that of the majority of "Joe Sixpack's" who base their concerns on Terminator, Bender or an obscure passage from an ancient religious text. In other words, those with knowledge of what is possible today are more informed about what that may lead to tomorrow. In the same vein I have an (old fashioned 1990-ish) degree in computer science which included an AI component, I have written numerous AI toys over the last 30yrs, and recently sat through the 2010 online MIT course* for AI (just for fun). None of this means that Joe is a moron, Joe is quite likely to be an "expert" in other fields compared to you or I.
Here's the rub with "AI" (or any complex and controversial issue). Due to a messy divorce I hadn't really been paying attention to what had been going on in AI during the 2000's, it blew me away, I showed my 'wife' who happens to be a (sought after) business 'expert' who's lectures attract large audiences. She shrugged and said "The voice thing is neat, but what's the big deal, it's just looking up the answers on the internet, right?". To this day she simply does not have sufficient knowledge to recognize the problem. Just about everyone I have shown (other than fellow AI geeks) has a similar reaction. Not only don't they "get it", they don't even recognize "it" when "it" is talking to them. There's no offense intended when such an opinion is deemed "uninformed", it's actually a plea for Joe to familiarise himself with the problem before offering an opinion.
Personally, I'm not afraid of AI suddenly turning hostile, but "knowledge is power" so I am definitely concerned about what the "known behavior of the only high intelligence we've ever met" may do with such a tool/weapon. Given the track record of our species I don't think that is an unreasonable concern, in fact the last line of your post would seem to agree with it.
Now, if you actually take a few moments to (randomly) listen to what these people are saying about AI in their speeches and interviews, you may find that their concerns are not that different to yours and mine and that the "SkyNet" hype is just the MSM doing their thing to "sex up" conservative (and unoriginal) speculation about the human tendency to use tools in every endeavor, including our inhumane endeavours.
* For anyone wanting to sharpen their existing AI knowledge (and make smarter toys), I highly recommend MIT's online AI course. It took me about a month to watch and absorb all the lectures, I approached it as a "refresher" but also learnt some new tricks that were not available 25yrs ago. The guy running the show has trouble keeping his pants hitched up but he is simultaneously entertaining, intelligent, and down to earth, when I finished the series I wanted to sit down and talk "cabbages and kings" with him...
And did you exchange a walk on part in the war for a lead role in a cage? - Pink Floyd.
I listened to a podcast (Econtalk, which is about as sober as podcasts get) that interviewed an AI "worrier" and he acknowledges that our current technology can't produce a superintelligence now. But he does make a couple of interesting points which I think make for reasonable discussion even if it isn't the "ZOMG, PANIC" kind of talk you imply.
One, discussing machine superintelligence before it actually develops is almost necessary because once it DOES exist it may be difficult to control. By definition, superintelligence will be smarter than we are and capable of manipulating at a level of complexity we can't grasp, making it hard to control.
And we've already created single-purpose "intelligences" similar to this, like the old "Internet worm" or some kinds of computer viruses that while they lack general purpose intelligence, have a self-replicating intelligence that can be difficult to contain. Imagine a smart hypervisor designed to manage a computing cluster but with the intelligence to replicate/migrate nodes across cloud computing infrastructure. Couple it with cyber defense technology, encryption, etc but given the single minded purpose of "don't shut down". It's not hard to see at a not-so-far-off level of intelligence that it could self-migrate across cloud platforms, resisting shutdown, possibly even able to hide in private cloud platforms all while being able to escape detection and control.
Which brings up the other point -- we don't know what machine superintelligence will look like. Part of the problem with understanding what superintelligence could be is that we don't know how far we are from creating one because as humans we try to imagine superintelligence in anthropomorphic terms using human epistemologies. It doesn't have to be the anthropomorphic HAL 9000, it could be a hypervisor manager, a securities trading system or some other single-purpose automation system that contains a feedback loop between a series of "dumb" systems coupled with a control plane. We may create it by accident and even if its not perfect, there are some realms where it wouldn't take long running amok to cause large problems, even if the outcome wasn't "judgement day".
"and only a slight majority said it would be a net positive."
So the proper headline is actually "Majority of actual AI researchers believe strong AI is good".
How do you build this 'idiot savant'? How do you give it 'motivation'? What do you tell it to work on? We are like stone-age man here, we don't even know about the atom and you want us to work on nuclear power? Its absurd.
But what is even MORE absurd is the idea that if we could create such a process that we wouldn't be in control of it. The very notion that somehow an intelligence could 'reach out of the computer' strikes me as the uttermost level of boogeyman type nonsense.
Here's a good analogy. You worry about 'making a more deadly virus' but we don't even HAVE a virus AT ALL right now. We would have to invent the very concept of a virus first, and we have zero idea about how to do that, beyond some incredibly vague hand-waving.
"Malo periculosam, libertatem quam quietam servitutem." -- Jefferson
The alien invasion nutters drive me crazy too! Imagine a race so powerful it can fling itself across the unimaginably vast gulf between the stars. What is Earth to such beings? There's nothing here you can't mine vastly more cheaply in some asteroid belt or on some moon somewhere, or out in the Oort Cloud, etc. With that kind of power you have no need to "find a place to live" etc.
In all of these cases it is very wooley thinking combined with some atavistic fears. Truthfully even the most clever humans rarely think rationally. People imagine that somehow a Stephen Hawking is some paragon of super rational logical thought or something, but its just utterly not true. These guys are VERY VERY clever and imaginative, but they don't really have any extra special insight. Neither do AI experts, apparently.
"Malo periculosam, libertatem quam quietam servitutem." -- Jefferson
'man' has long been an existential threat to 'himself' and all living things on earth... which would his creation not also be such a risk?
The sense of self is illusory. Our hopes/dreams/desires/proclivities/preferences are largely the result of conditioning. Our judgements are hopelessly distorted as if in a funhouse mirror.
Yes, AI will disclose how biased and foolish we are and how really really stupid. That's an existential pretty big deal.
It's the mouthpiece of the Falun Gong. Someone needs to challenge the Wikipedia article about it (http://en.wikipedia.org/wiki/The_Epoch_Times).
Not to be trusted.
An expert is a drip under pressure, also anyone 200 miles from home and carrying a briefcase.
Um..that source the epoch Times is so lacking in any scientific understanding, paranoid and distorted views of this world that I would take even their weather report with a drop of soy..it is funded by a religion ..enough said.
If strong AI was an existential threat then it would already exist in the universe, and would have gobbled us already if it cared to.
If AI reached human levels then it would be subject to the same existential choices that we have.
because that AI has embraced the absurdity of the universe and is capable of all the insanity of man.
Flag as Inappropriate
obat benjolan di gusi
obat gondok anak
obat ispa anak
The people polled have a financial interest in AI being thought to be safe. And despite that, many of them are scared.
I'm scared, too.
All of those claims about what AI research is and isn't are wrong.
Even if they were true, they would only represent the present, not the future. People were so sure the Wright brothers couldn't fly ...
The Ebola virus can't do those things. So you're not scared of it, right?
Just because it's not like a human, that doesn't mean it can't kill you.
until an electronic Nietzsche proclaims root is dead.
Of course, many nuclear scientists didn't think their atomic bombs would be a plus, but not being ready to combat the Nazis was apparently a big motivator. So why are these 13 or 18 percent pushing something they think is bad? Just too cool a subject to leave alone?
I'm more afraid of human intelligence. Let's see: war, overpopulation, eco-destruction.... what more could AI do than we have to destroy our planetary ecosystem?
The Federal government is unlikely to ever put itself in a subordinate position to any computer system because computer algorithms don't like bad bookkeeping. Our government has all the worst attributes of an evil AI and none of the disadvantages for those humans at the top who seek power and money.
This is absolutely true, and is related to the whole "what is intelligence?" question. We tend to equate it with intent and free will because we're used to dealing with humans or at least animals that have a survival instinct and thus evidence 'will' to go with it.
There's no reason to believe that artificial 'intelligence' will ever evidence the same sort of intent. There's very little reason to endow a stock trading application with a sense of self or a need to survive. Even if it had some rudimentary form of these things, for whatever reason, they need not take the form of concern about itself, it would be more likely that such a program would be designed to protect its funds!
I can imagine some far future where in some sense self-driving cars have a 'will to survive' that makes them 'want' to avoid crashes, but they're unlikely to have the kind of self-awareness or ability to generalize that would be required to turn that into a desire to do anything except avoid traffic accidents, which is all they will need to understand about the world. You could imagine an interstellar space probe built to the level of a full general AI simply on the basis of having virtually no idea of what it will run into, but we're centuries from being able to build such things. As long as it can phone home fairly quickly all it might need is self-driving-car level 'reflexes', that would be useful. Even Pluto Express gets by as a totally dumb remotely controlled instrument, as do the Mars rovers. They don't need to be true AIs, not even close.
"Malo periculosam, libertatem quam quietam servitutem." -- Jefferson
Sure. Did it to myself decades ago. Offspring of my genetic line aren't of the least bit of interest to me; perfectly happy raising kids of other birth who needed parents (5 so far, mostly excellent results.) Plus that whole "all the bareback sex with my SO we want, any time" thing is awesome.
Which, again, is just how I approach feline guardianship. Don't need new kittens from them. Plenty of kittens out there that need to own their own human.
I've fallen off your lawn, and I can't get up.
The responding subjects were dominated by people who attended conferences on AI such as “Impacts and Risks of Artificial General Intelligence”. If you look at the published researcher groups: 60% are positive and only 8% think it will be extremely bad. The 2 negative choices were 'On balance bad' and 'Extremely bad (existential catastrophe)', so its not fair to say they specificly endorsed 'catastrophe', just a stronger option than 'somewhat bad'.
Instead of "170 of the leading experts" it should say "170 people who went to conferences on AI Risk"