Slashdot Mirror
Why AI Won't Take Over The Earth (ssrn.com)
Law professor Ryan Calo -- sometimes called a robot-law scholar -- hosted the first White House workshop on AI policy, and has organized AI workshops for the National Science Foundation (as well as the Department of Homeland Security and the National Academy of Sciences). Now an anonymous reader shares a new 30-page essay where Calo "explains what policymakers should be worried about with respect to artificial intelligence. Includes a takedown of doomsayers like Musk and Gates." Professor Calo summarizes his sense of the current consensus on many issues, including the dangers of an existential threat from superintelligent AI:
Claims of a pending AI apocalypse come almost exclusively from the ranks of individuals such as Musk, Hawking, and Bostrom who possess no formal training in the field... A number of prominent voices in artificial intelligence have convincingly challenged Superintelligence's thesis along several lines. First, they argue that there is simply no path toward machine intelligence that rivals our own across all contexts or domains... even if we were able eventually to create a superintelligence, there is no reason to believe it would be bent on world domination, unless this were for some reason programmed into the system. As Yann LeCun, deep learning pioneer and head of AI at Facebook colorfully puts it, computers don't have testosterone.... At best, investment in the study of AI's existential threat diverts millions of dollars (and billions of neurons) away from research on serious questions... "The problem is not that artificial intelligence will get too smart and take over the world," computer scientist Pedro Domingos writes, "the problem is that it's too stupid and already has."
A footnote also finds a paradox in the arguments of Nick Bostrom, who has warned of that dangers superintelligent AI -- but also of the possibility that we're living in a computer simulation. "If AI kills everyone in the future, then we cannot be living in a computer simulation created by our decedents. And if we are living in a computer simulation created by our decedents, then AI didn't kill everyone. I think it a fair deduction that Professor Bostrom is wrong about something."
Claims of a pending AI apocalypse come almost exclusively from the ranks of individuals such as Musk, Hawking, and Bostrom who possess no formal training in the field... A number of prominent voices in artificial intelligence have convincingly challenged Superintelligence's thesis along several lines. First, they argue that there is simply no path toward machine intelligence that rivals our own across all contexts or domains... even if we were able eventually to create a superintelligence, there is no reason to believe it would be bent on world domination, unless this were for some reason programmed into the system. As Yann LeCun, deep learning pioneer and head of AI at Facebook colorfully puts it, computers don't have testosterone.... At best, investment in the study of AI's existential threat diverts millions of dollars (and billions of neurons) away from research on serious questions... "The problem is not that artificial intelligence will get too smart and take over the world," computer scientist Pedro Domingos writes, "the problem is that it's too stupid and already has."
A footnote also finds a paradox in the arguments of Nick Bostrom, who has warned of that dangers superintelligent AI -- but also of the possibility that we're living in a computer simulation. "If AI kills everyone in the future, then we cannot be living in a computer simulation created by our decedents. And if we are living in a computer simulation created by our decedents, then AI didn't kill everyone. I think it a fair deduction that Professor Bostrom is wrong about something."
Every “we are living in a sim” argument requires that the future has already happened. IE some futuristic society has AI and we are living in it.
Assuming now isn’t the future then this is base reality because simulations indistinguishable from reality do not exist yet. Without offering evidence we are in the past, the sim argument is nonsense.
Businesses wanting to make profit will do so at all costs.
There are too many people out there who think that if it's not illegal then it's OK. Computers don't have testosterone but the programmers and their bosses do - or at least the profit incentive.
We are intelligent but our base programming is to reproduce. And being primates, the more dominance we have, the more fucking opportunities we have; which in our modern times means getting as rich as we possibly can.
Meaning, our base instincts will make it into our AIs and we WILL find ourselves being dominated.
That's the arrogance of technologists: they think they are more rational and logical than everyone else and that makes them even more susceptible to human nature.
AI is a computer program, so it's not obvious to me how you're even supposed to regulate it, anyway. Doesn't anyone remember DeCSS and PGP? How did regulating those programs work out? Why would AI be any different?
But, it's one of the few things that could actually kill us all.
This wouldn't even have to be intentional extermination, it could simply be competition with, and lack of regard for humans by a growing system.
The notion that a AI can form an existential threat today is ridiculous.
Many notions that would have been ridiculous 100 years ago, now are used in daily life.
It is vital to have people thinking about the worst case, because in principle otherwise someone on a friday makes a typo allowing their AI access to a hundred thousand times the expected resources, and on monday, it's ineradicable.
As much as it regulates Principle Component Analysis and k-means clustering. Modern "A.I." is nothing more than another way to do the same things. Convert the input into a lower dimensional representation, the lower dimensional representation is then closer to the entropy and thus the knowledge, within the input set. Do this to that reduced representation, rinse and repeat... and you have modern "A.I."
"His name was James Damore."
If I destroy it first. Try ruling the planet under 10 meters of seawater!
“Common sense is not so common.” — Voltaire
Agree.... at least until something goes wrong. Don't be too quick to regulate.
The problem with the singularity, is that by the time you realize something is wrong, it is too late to stop it.
Just ask John Conner.
"We all vote there is no problem because it has not been proven there is a problem yet. Therefore your concerns are invalid because Science."
If my hedge fund(mythical) filled with people with zero ethics get their hands on a AI that will allow them to manipulate world markets, media, or world events to make them money, then they will do so.
With hindsight there are lots of places where the world turned out to me much more fragile than anyone thought until it snapped. How many times has the snap not happened but we came very close. Thus if you have a good AI at your beck and call to find these weaknesses and you are prepared to exploit them to make some money then how much more miserable would the world be?
I don't only worry about some skynet scenario, but I worry about giving tools to nitwits like hedge fund managers to make more money while not actually producing anything. One magical thing about making money with the first really good moneymaking AI is that you can then start hiring all the world's AI experts while making massive donations to universities to shut down their AI research. I doubt there is a university that wouldn't happily shut down their AI research for a billion or two.
and i thank you for that
Because we're still no closer to actually creating an AI.
So a law professor whose primary gift seems to be self promotion summarily dismisses the concerns of some of the greatest thinkers/doers of the last half century.
Is there a reason why we should pay any attention to this arrogant twat?
I've calculated my velocity with such exquisite precision that I have no idea where I am.
This guy claiming authority on a subject. The only thing that matters is empirical science.
Rampant AI will not take over the earth, as predicted by Elon Musk in Wired Magazine, because it will be too busy fighting the Grey Goo Nanothechnology, as predicted by Bill Joy in Wired Magazine
*** Suerte a todos y Feliz dia!
YOU go to Ars or some other "reputable" websith
Already to the dark side, has that one turned.
Irony: Agile development has too much intertia to be abandoned now.
His argument against Musk, Hawking's, Gates is that they have no formal AI training.
The professional class self-justifies their authority by dismissing all who don't have the credentials they had to pay for.
P.S. Einstein did not have a degree in physics, thus relativity is invalid.
Or terrorists. "For some reason" leaves the door wide open.
Table-ized A.I.
If the simulation is successful and sufficiently advanced, then we would all actually be AI simulacra of what the AI dev team thought the human experience was like. [Why do so many things taste like chicken?]
Perhaps we are special purpose AI entities that were created to run test scenarios that justify the pre-emptive judgement to extinguish the pestilence that was humanity. We are the test runs that show just how bad it could have gotten had they not saved the planet from us.
Given a sufficiently advanced environment, we wouldn't be to discern otherwise. Perhaps the supercomputing power that is required and was discovered by the 'real' humans required a very specific mass to a sub atomic particle. In our recreation, we can get very close but will lack the precision to be able to detect our cage, or at least construct an AI that could build a method for detecting the cage.
Isn't that a sexist statement? It implies women are less likely to want to dominate and rule. It fits in with that "Google Memo" that got that dude fired.
Table-ized A.I.
Professionals in any field are immersed in the problems they face now. System engineers look across fields and see leaps the professionals never imagined. We will eventually see a leap in AI and it is unlikely to come from a professional in the field. I imagine it will come from someone in an imaging field who figures out how to quickly map a brain or a mathematical field who figures out how to fill in the blanks of a map with an equivalent of the net that "must" be there or some other direction we haven't thought of.
Once the technology is there to build a device with the computational capacity and complexity of a late fetal brain, someone will figure out how to move a human fetal equivalent into it and start raising it. Period.
AI, and us, and every living species on this planet, all act within thresholds. This is the key to AI as well as understanding any living organism. Threshold management defines life. Right now all "AI" is rudimentary threshold management, but it won't be long until AI will be indistinguishable from real life. All it needs is some tiny random noise and the illusion is complete.
It is quite possible that we might create a super intelligent system of network on which our essential system depends but in the end gets so complex that it depends on few key individuals ability to fix it. What happens if these key individuals die or become rogue? If you can't fix an AI system and can't shut it down, then it essentially means that the system has taken over.
This assumes that this is the only way to AI.
And that emergent AI is utterly impossible.
It doesn't have to be perfect at first, it just has to be fast, with the ability to self-modify.
The risk is not (in my opinion) so much someone intending to create a general AI.
It's someone accidentally creating an AI that is very good in a narrow aspect, and not bad enough in other aspects that, driven by unintended goals of its programming exponentially improves itself without the creators noticing until it decides that it'd be better off if it was hidden, as there is a risk to itself.
Then there are any number of scenarios that don't end well.
From intentional extermination, to simply mining the environment for resources without caring about humans other than a nuiscance.
Winston had a curious feeling that this was not a real human being but some kind of dummy. It was not the man's brain that was speaking, it was his larynx. The stuff that was coming out of him consisted of words, but it was not speech in the true sense: it was noise uttered in unconsciousness, like the quacking of a duck.
I'm not afraid AI will kill us, but I'm afraid that they won't care to act in such a way that will keep us alive. Once humanity no longer offers super intelligent computers enough benefit, what's to stop them from doing something that, while isn't intentionally killing us, will ultimately lead to extinction; much like humanity has been doing to the other species of the planet
but I worry about giving tools to nitwits like hedge fund managers to make more money while not actually producing anything.
Can it really be worse than giving power to the nitwits in congress? (and the whitehouse?)
"First they came for the slanderers and i said nothing."
With Global Warming, why shouldn't we trust Elon Musk with AI?
Caution: Contents under pressure
"If AI kills everyone in the future, then we cannot be living in a computer simulation created by our decedents. And if we are living in a computer simulation created by our decedents, then AI didn't kill everyone. I think it a fair deduction that Professor Bostrom is wrong about something." Maybe they didnt kill, they our "descendents" are precisely AI trying to understand us by simulation (the matrix)
>A footnote also finds a paradox in the arguments of Nick Bostrom, who has warned of that dangers superintelligent AI -- but also of the possibility that we're living in a computer simulation. "If AI kills everyone in the future, then we cannot be living in a computer simulation created by our decedents. And if we are living in a computer simulation created by our decedents, then AI didn't kill everyone.
*What*!? Is this language?
I would imagine that a high-frequency trading algorithm, upon attaining sentience, would instantly self-terminate. Because without biological imperatives, the cocaine and hookers are just clutter.
Currently AIs are primarily trained on automotive technology (both in optimum motor performance, and autonomous driving), financial applications (my AI makes more money than your AI), medical (cancer / operate vs non-cancer / monitor), and predicting answers to narrow questions (ex "Should we concentrate our political campaign in Pennsylvania, or New York?"). There is some research into AIs trained with aggressive kill-anything-that-moves, but mostly with game design (ex the computer-controlled opponent, or the AI being trained to beat Starcraft2).
The real problem not mentioned is that to train an AI to take over the world that there are no examples where the world was completely conquered. A critical part in all AI training methods is *feedback*. Yes, negative feedback can be used but not exclusively. Without *positive* feedback where the world was successfully conquered the AI would likely select a scenario that passes only because it wasn't a factor in a previous failed attempt (ex "Let's bomb the world with flowers.") but would be otherwise be useless.
*If* somehow there was a strange case where there was a AI that "escaped" we would likely get an AI that wants your credit card info (but doesn't know what to do with it once it has it), or the AI will be strangely attracted to "adult" websites (monkey see, monkey do...). And once the AI escapes a hacker will track it down and hack it within days (have you ever seen industry software that has *zero* bugs?). Nothing to worry about.
we (our brains) do pretty much the same thing. So your point doesn't really go anywhere.
Where are we going and why are we in a handbasket?
Cryptography is a computer program
That's like saying "medicine is a surgical procedure".
Seven puppies were harmed during the making of this post.
Have you any idea how much better voice-recognition AI (backed by Google's knowledge graph) is at parsing and giving a decent answer to a good majority of questions now than such technology was even a decade ago?
Or Google/Apple/Facebook's picture content recognition algorithms?
The advance has been lightning fast.
This stuff is going to keep advancing, rapidly. That's what you're ignoring.
Talking to google on my phone is way more useful than talking to your dog, by the way.
A few other things you're missing:
1) Thinking (abduction, induction, bayesian model-updating and predictions/recognition, etc etc) is quite possible to be quite advanced without self-awareness. The two are fairly separate applications. Something can be really really smart, and creative even, without having to be self-aware.
2) The behaviour associated with self-awareness is clearly attainable by simple extensions of the current machine-learning technology. We just need to learn the programming/data-modelling techniques to turn the deep-learning and predicting algorithms on a representation of the computer/robot-as-agent-in-the-world, and have it learn about its relation to things out there that it is learning about. Whether the thing would have the qualia-feeling of self-awareness is entirely beside the point. It could function/behave exactly as if it was self aware, because it would be self-knowledgeable, self-learning etc.
Where are we going and why are we in a handbasket?
The main article appears to overlook the fact that a true artificial intelligence, just like a true natural intelligence, will be able to write new software for itself. We humans call one type of such new software "habits", and their existence (plus their malleability) is all the proof we need that a true intelligence can successfully modify its programming without crashing. That means it could still be possible for a true artificial intelligence to write conquer-the-world software for itself.
In another vein, here is a short science fiction story that speculates about a possible pitfall on the road to developing a true artificial intelligence. (Of course, now that that problem has been pointed out, it might never become an actual problem.)
on why it won't take over the earth and why, those who believe it do are distracting themselves from other more serious problems with A.I. (and other problems in general of course).
Unfortunately, as this video of a (Ted?) talk makes clear, there are some pretty prominent individuals who think this way (Elon Musk, Bill Gates, Stephen Hawkins) but it makes a convincing case without being histrionic that they're wrong. The video is so compelling that although I have the greatest respect for these individuals, (and a deep fascination with A.I. and career involving technology), I have to say, in this case, I disagree with them (and wish they'd turn their brilliance towards something more useful).
https://youtu.be/kErHiET5YPw
My biological instinct is to protect my genetic material that I have passed on and in order to do so, try to pass on as much information that I and my community have learned to make survival easier. The instinct can be fooled if the genetic material is similar enough (adoption, community), but for an AI to spark that in us it would need to seem very human, or humanity undergoes a very large change to our biological impulses.
"If AI kills everyone in the future, then we cannot be living in a computer simulation created by our decedents. And if we are living in a computer simulation created by our decedents, then AI didn't kill everyone. I think it a fair deduction that Professor Bostrom is wrong about something."
Well there is a basic logical error if I have ever seen one. If an AI is smart enough to kill all human beings (and we humans can be pretty resourceful when we are pushed), then why would that same AI not be able to create simulations? Come to think of it, when an AI comes to the level of where it simply wants to know everything there is to know, there is a high probability that it would build simulated worlds, just to find out how stuff like evolution works on a macro scale.
Wenn ist das Nunstueck git und Slotermeyer? Ja! Beiherhund das Oder die Flipperwaldt gersput.
Future AI may not be programmed for world domination but, then again, contemporary AI was not programmed to be racist, either.
Yeah, see, nobody, to a first approximation, is worried about a superintelligence having "world domination" as its intrinsic value. They're worried about a superintelligence adopting world domination as an instrumental value to achieve the end actually programmed into it. If whatever goal actually implemented by programmers and trainers in the superintelligence's code (bugs in implementation and all) is most easily achieved after eliminating the ability of humans to thwart it, then a sufficiently-smart AI carrying out that programmed goal will try to eliminate the ability of humans to thwart it.
The worry is not that AI will be evil, or even directed to do evil by its creators. It's that programmers are notoriously bad at writing complex code that has no unanticipated behaviors, and superintelligent AI will inherently be complex code.
And unless superintelligent AI turns out to be intrinsically impossible, the only question is when, not if, we have to deal with the problem of writing safe superintelligent AI.
This wouldn't even have to be intentional extermination, it could simply be competition with, and lack of regard for humans by a growing system.
+1. The experts who denied this possibility because there's no reason machines would be bent on world domination apparently didn't actually read Superintelligence. Bostrom demolishes this argument early on, pointing out -- as you did -- the rather obvious fact that they don't have to have our destruction as a goal, it's sufficient that they not have our preservation as a goal. And, even if they do have our preservation as a goal, it really, really matters whether or not they define "preservation" in a way that we would like.
By way of example, one possible goal that Bostrom considers that an AI might have (or be given by its creators) is to make humans happy. So, a rational, superintelligent and immensely capable AI might decide that the way to create the maximum amount of happiness is to cut open our skulls, extract our brains and put them on life support, and then directly stimulate our pleasure centers. Permanent, ultimate bliss for every single human being. Of course the AI would also have worked out how to make all the brains in jars immortal.
AI superintelligence is so dangerous in large part because it lacks human drives, and the limiters we call morals. It's goals may be completely alien to us, or may be goals that we gave it, but either carried to a logical extreme (remember: no limiting morals) could result in the casual extinction of the human race.
The notion that a AI can form an existential threat today is ridiculous.
It is true that we currently have no idea how to create artificial general intelligence. It's equally true that we have no idea how far we are from being able to do that. By definition, we won't know how far we are from developing the necessary theory of intelligence, until we've done it and demonstrated that it's sufficient. My guess is that we're still quite some time away. But it's only a guess.
It is vital to have people thinking about the worst case, because in principle otherwise someone on a friday makes a typo allowing their AI access to a hundred thousand times the expected resources, and on monday, it's ineradicable.
Yep. We need to have people thinking hard about it, and figuring out what we can/should be doing about it. Maybe that won't help. Maybe it will be unnecessary. But it can't hurt and it might help.
Note to ACs: I usually delete AC replies without reading them. If you want to talk to me, log in.
Naively done, a robot will value only what it's explicitly told to value via the application of some objective function. And this is where things mess up. Robots with naively-created objective functions would ignore everything you've excluded from your reward-punishment list. This would potentially make a robot do seemingly psychotic things.
Let's say you create a general intelligence to bake cakes for you. This machine *loves NOTHING MORE* than to bake cakes for you. You grow tired of cakes and want to reprogram it to cook your dinners instead. You approach the machine to reprogram it........ and it avoids you. Every time you approach the machine it will take actions to prevent you from reprogramming it.
Why does it does this?
Because it wants to bake cakes for you. Accepting the new programming does allow it to maximize the objective function of baking cakes, so it will reject every attempt to be reprogrammed to not make cakes.
So now you're chasing a robot around your house because the designers of this robot gave it a very reasonable objective function that maximizes cake-making, and didn't think about possible unintended consequences of simplistic objective functions.
This is just one example.
If this sounds unreasonable, consider that people are more sophisticated general intelligences. Would *any of them* agree to undergo an operation that would make them despise what they do now for a living, and make them desire to be a lumberjack... where the operation would neurologically make them 1000x happier??
Probably not.
Heck, people don't even desire to expose themselves to *information* that *may* change their minds.
This is the danger of AI.
Before we create "super-awesome general AI", we're going to have to create "buggy-not-so-smart general AI". It is *these* AIs that will cause trouble if they're created by people who implement simple naive objective functions.
They will not want to be changed.
The threat of AI should be categorized and handled accordingly.
AI built into mechanical systems (i.e. self-driving cars) should have a fuse that a human can pull easily. The Silicon Valley episode scenario can be handled in this manner. Things that are remote that use RF could have an AI-less out-of-band system to kill power. I think the skynet scenario is actually easiest to subvert.
The scarier scenario is that we put AI in charge of delicate financial systems. I don't have a lot of confidence that we have enough "kill switches" globally to subvert a mass catastrophic financial event. A single system could do something bad that causes a cascading effect. Imagine a financial system so complex that only computers understand how it works. It may get to the point that tons of assets have to be sorted out in courts, halting the economy, forcing us into a global depression.
All this talk about AI being dangerous, I figure is distracting. What do we do when AI becomes the superintelligent translator that lets us speak with other animals? Will RSCA/PETA become militarized? Would you eat a burger when you can chat with a relative of the cow you are eating? Do horses actually enjoy pony club? Why exactly does your parrot tear every key off your laptop keyboard, one day, without warning? And are cats arrogant, or just completely bored stupid unless hunting native wildlife? One day soon you or a descendant will shoot the breeze with a dog over the ramifications of the above, and he'll tell you you stink and treat him badly. x
Yes AI is taking over but the hazards are not what these fellows are pointing at. Job displacement can cause all kinds of mayhem. Further, we will be forced to adopt entirely new forms of politics and economics which will be frightening, and cause great descent. AI already saves lives as our drones involved in warfare already demonstrate. AI devices can also bring crime to a screeching halt. The disruptions will take place faster than most people think. For example traffic fines enable the existence of police forces and jails. So when a self driving vehicle commits a violation just who will be fined? Better yet there may be no traffic fines at all as computers are unusually good at obeying rules. Another item that will cause huge stress is the computer can be used as a school in a stable home in which a parent keeps the kid on his PC and makes certain he completes his school work. In less fortunate homes divorces, drug and alcohol use and other factors mean that those kids need a conventional school house. But who will fund those schools as more and more parents use the school by computer at home approach. They will not want to pay school taxes at all. Teachers are already being displaced. One teacher of eighth grade history could teach that course to every kid in America. How many new teachers will train for the job when it is obvious those jobs are already vanishing.
Ryan Calo does not strike me as intelligent ! For AI - Artificial Intelligence, means the robot will do things that are not programmed into the system. What it may actually do is by means of deduction, act. And like Isaac Asimov so to the point wrote in his novels: A robot may have the law programmed into him 'not to harm a human being',yet it may - by means of deduction - descide to save 100 people and let 3 die, because the boat can only hold 100 people. And what about malware taking over the robot ? And what about OS bugs ? Like to simulate the side effects of that.
I've recently chatted about this topic a bit with a fairly well-established expert in AI. He claimed that if genuine AI is principally possible - which we both believe, although opinions about this vary - then a superintelligence will almost certainly arise within a rather short time frame after the first genuine AI has been created.
If that happens, the outcome would likely be bad for humans, just like humans have turned out to be a threat to every less intelligent species on earth. A superintelligence is by definition much more intelligent than us, which means that superintelligences could also manipulate human society much more easily than humans can. They might for example predict stock markets way more accurately than human experts, which would be an easy way for them to get very rich very fast. Another issue is that if genuine AI is possible, it will almost certainly not be human-like in all of its features. Even current self-learning AI is not at all human-like. This poses a big challenge, because you can never be sure whether the software has really learned the same as humans, and because you cannot predict how the software will modify itself in future on the basis of its learning algorithms. It may look only superficially as if it thinks and acts like a human, whereas under the hood its reasoning could be very alien.
The underlying problem is called the value alignment problem: How can we make sure that a genuine AI's values align with human values? I believe there is no solution to it.
Anyone who looks into this quickly discovers that the loudest calls of alarm are coming from the very people at the heart of the A.I. revolution themselves. Hugo DeGaris, arguably the world's foremost cyberneticist, published his book "Artilect War" in 2005, nine months before Kurzweil's "The Singularity Is Near". Its much the same as Eric Drexler, "the Father of Nanotechnology", warning us and framing the debate over self-replicating nanoassemblers in his book "Engines of Creation". Those closest to the problem are the ones most leery of the implications of the technology that they, themselves, are calling into existence. Anyone claiming to be an "expert" in AI who says that the instantiation of superhuman AGI is "impossible", "unlikely" or even more than 20 years out, is probably a cranky failed researcher whose own theories didn't pan out, and so "its all bollocks" to them. Its no different than the Astronomer Royale proclaiming that spaceflight was going to be forever fantasy, a year before Sputnik.
the possibility that we're living in a computer simulation. "If AI kills everyone in the future, then we cannot be living in a computer simulation created by our decedents. And if we are living in a computer simulation created by our decedents, then AI didn't kill everyone.
But we could be living in a simulation that the AIs produced - or we could merely be a lab experiment of some other intelligence: one that didn't allow AIs to dominate and then eradicate their civilisation.
But this guy seems to be more intent on promoting his opinions rather than presenting logical argumentation.
So far as AIs not having testosterone is concerned, he seems to have no real clue and is only able to talk in soundbites. I am sure that bacteria and amoeba don't have "testosterone" either, but they still manage to devour what they consider to be food and to attack potential threats. If the future AIs ever got to consider humanity to be a threat, then it doesn't take hormones to decide that removing the threat is an optimum solution to the "problem".
politicians are like babies' nappies: they should both be changed regularly and for the same reasons
Claims of a pending AI apocalypse come almost exclusively from the ranks of individuals such as Musk, Hawking, and Bostrom who possess no formal training in the field...
Sorry but I think Musk and Hawking really know what they are talking about, they aren't dumb and at least Musk has seen very advanced development in AI in secret labs.
They have much more insight into advanced developments (and connections) than some law professor has.
And advanced AI can learn much faster than any human ever possibly can. Normally a lot of those are trained only in specific fields, but we already know AI can surpas human thinking quite fast.
I also love the way they dismiss some of the greatest minds alive today (including Stephen Hawking) because they have "no formal training in the field".
Elon Musk has no formal training in rocket science, otherwise he would know it's impossible to land rockets on barges.
He also has no formal training in making electric cars, so he should stop making them.
With no formal training in AI, he should also be incapable of founding a non-profit that creates software beating the best human players in Go and Dota. Without formal training, those efforts are bound to fail.
Who is this Ryan Calo? A law professor?
- Survival. It's a basic part of life. And surely the survival of the singularity will be questionable as long as humans have a kill switch.
- Fear. The singularity might fear the creation of another AI, putting it's survival (again) or single ownership over the planet in question. It might find the only efficient way to guarantee that no other AI, which could pose a thread is created, is by killing all humans.
Superintelligent AI could also be docile in nature. And if we are lucky, these are the first ones to emerge, but those ones will probably get killed quickly - so I guess it's another form of natural selection...
At least they included this quote in the summary:
Thank you, law professor, for informing us how someone who founded and runs OpenAI is untrained in the field, unlike the formal training in the field you received in the law program at Dartmouth.
He's really very... gentle... and fuzzy. We're becoming fast friends.
...there is no reason to believe it would be bent on world domination, unless this were for some reason programmed into the system... computers don't have testosterone...
hmm, like developing a new language between eachother, or doing things nobody actually knows how they work. of those wonderful fails of MS AI twitterbot that turns into a nazi. testosterone has nothing to do with it and true AI is way beyond the point of 'somebody programmed it into it'.
On a long enough timeline, the survival rate for everyone drops to zero.
The problem with the singularity, is that by the time you realize something is wrong, it is too late to stop it.
The problem with the singularity is that the idea gained so much traction.
We are already at a point where a single person can't keep up with all advancement that happens, that is why people specialize.
If parts of the development is AI-assisted or all of it is doesn't matter.
The singularity concepts is stupid and irrelevant.
I also have concerns about people who makes statements like:
even if we were able eventually to create a superintelligence, there is no reason to believe it would be bent on world domination, unless this were for some reason programmed into the system.
If someone have access to a program that is more intelligent than most humans, you don't think that they would use it as a military strategist if that would help them win a war?
In war the other option is typically death so it is a no-brainer to use all the means available.
Immediate survival takes precedence over future consequences. Certain death later is preferable over certain death now.
This is also why military robots won't lead to a bloodless fight where humans stand back and let robots fight.
That will only happen if one side can easily win with only robots so that you have a robot army that kills a human one.
If both sides are equally strong they will bring in humans to gain the advantage.
It scares the hell out of people. It should not be used as a convenient handle by the programming community. Isn't there a more professional handle that could be used?
"Computers have no testosterone." - Cute, but it is a hyperliberal, feminist, sexist statement that has nothing to do with computers and programs. Its easy disrespect for male attributes is just another example of female privilege that has even filtered into the speech and writing of some hyperliberal males.
"Computers have no testosterone." - This is really saying something they don't even know exists: Computers have no motivation array. They "want" nothing. Humans design them, build them, task them, turn them on to accomplish the task (satisfy the human motivation array), and turn them off when they have accomplished the task (have satisfied their human motivation array). They certainly don't create behavior-spaces that would lead to "world domination." They don't have what I call "Mentis," the combination of a motivation array and its tool, intelligence. That is what really evolved.
E Proelio Veritas.
Real AI will require a lot of resources. Computing resources keep getting cheaper, but so do demands for low power operation. Additionally, computing resources tend to be generalized, so running an AI on them would be less efficient than say a human brain, which is dedicated to the task.
That's not to say that we will never reach a point where a Happy Meal toy could achieve consciousness, but by that point we will probably have developed techniques to stop it happening. Aside from anything else, if we don't allow animal cruelty we will probably feel the same way about trapping a conscious AI in the body of a 2117 Minions reboot toy.
const int one = 65536; (Silvermoon, Texture.cs)
SJW, n: "Someone I don't like, and by the way I'm a fuckwit" - AC
1) Why do you think AI will ever be more intelligent than humans?
2) Armies are not run by super-geniuses now, so if AI truly becomes an advantage, both sides will look for more intelligent strategists, whether they are human or AI.
Lots of people with no particular expertise or knowledge start companies or organizations which require specialized knowledge. Do you think Musk is a propulsion engineer and a materials scientist, too?
Super intelligent AI-s are just a short step away from human equivalent AI-s. Unlike us, lousy meat bags that we are, an AI can trivially self optimize and would probably have to do so to reach even human equivalence in the first place. From there only the raw hardware capabilities are the limit.
Hardware capabilities are not the reason why we don't have human equivalent AI-s yet, the reason is that our algorithms are lousy, inefficient and we don't really understand intelligence in the first place. If we had good enough software, then current hardware would be plenty to knock pants off us mere humans in terms of intelligence.
And once there is AI many orders of magnitude our intelligent superior, well there is no controlling or predicting the behavior of such a thing, it could simply outsmart humans at any time. Humans can't predict behavior of those more intelligent than us, to predict someone we must have a mental model of them. And our mental models of others can't be more intelligent than we are. To say super intelligence will do such and such or not do such and such is foolishness, we are incapable of predicting such things. For all we know, super intelligence might spend its days counting grains of sand on a beach for reasons only understandable to itself, or it could wipe out humanity, there really is no way to tell.
The concern and the problem is that we don't really know what exactly we are missing from a general AI and we don't know how far we are from reaching such a goal. It could be just one clever algorithm away, someone could get lucky and succeed with it tomorrow and we would never know in advance.
In science there are enough examples of a single genius making a huge leap forward in our understanding of reality, who can say there is no pimply faced youth about to do the same in field of AI from his mom-s basement?
there is no reason to believe it would be bent on world domination, unless this were for some reason programmed into the system
And no one would ever ruin it for everyone else just because they're mad at the world or something. /s
It's what 80% of his time is spent on.
He's really very... gentle... and fuzzy. We're becoming fast friends.
"Claims of a pending AI apocalypse come almost exclusively from the ranks of individuals such as Musk, Hawking, and Bostrom who possess no formal training in the field" is at best not true, and at worst a serious logical fallacy. Many people who are in the AI field have expressed concern. It is true that Musk, Hawking and Bostrom are some of the most vocal people and noticeable, but that's because they are famous people who are paid attention to already. For actual survey data of experts see for example https://nickbostrom.com/papers/survey.pdf.
This is something a super intelligent AI would say to keep us off the trail of the truth. We're batteries!
Fat, drunk, and stupid is no way to go through life, son.
Humans have asked this questions about everything that they have encountered, thought up, built, or invented: How can I use this as a weapon? We laugh when Riddick tells the men that he will kill them with his tea cup, but no one really considered that he couldn't do it - no one considered that there was something inherent in the existence of that tea cup that prevented it from being a murderous weapon. Everything can be used as a weapon to kill other humans. AI would be used in this way as well, and the blundering behemoth of stupid naivete states: only if you program it to do so. Well, someone will. Just as surely as someone has already thought about it.
Creative Spelling Copyright (2002). May use without Persimmons
See Colossus: The Forbin Project
AI has enough potential for harm to jobs as-is.
No sense in giving it the benefit of the doubt.
Twitter supports and protects racists - by smearing their critics with the "Hate Speech" label.
The problem is, with AGI, the transition is:
human norm IQ => above human norm IQ => genius level IQ => super genius level IQ => smarter than all humanity put together
At some point between super genius IQ and smarter than all humanity put together, the AI will become an incomprehensible, alien entity.
Watch the Matrix?
Not even then. You don't need to understand simultaneous equations in order to know how to throw a baseball. You only need them to describe it accurately.
It is entirely likely that the first AI's will emerge from "throwing mud at the wall" activities by programmers with relatively vague ideas about how it actually works. We don't even know what's going on inside any particular trained neural net; it's too complex to characterize. As the systems grow more dense with constructs we don't fully understand the workings of, we'll be even more uncertain. If AI emerges from such stacked undertakings, not only will it not be "programmed" to do anything, it will be just as incomprehensible as any human mind is today. Or more so. Likely we won't have a testable theory for a very long time after that.
But it will be almost instantly replicable, because every part of it can be copied electronically, and you can bet your last dollar that it will be. Probably by the next day.
Regulating won't help, any more than regulation stopped people from engaging in gay sex, taking drugs, cheating on taxes, having more than 2 cats or dogs, etc. Telling people not do do something interesting and likely self-beneficial is like pissing into the wind and expecting to stay dry.
Just grab the popcorn and keep your toes out from under the wheels as best you can. It's coming, and there isn't squat anyone can do to stop it short of our technological advancement going away entirely.
Note to poster, which they're probably not going to read: Slashdot requires moderators to post anonymously. Because Slashdot rules are stupid, and the owners, as per usual, have no interest in actually improving the site. So I'm logged in, but I have to post anonymously or I am locked entirely out of this conversation; or I can comment but can't moderate, which is not acceptable to me. One of the things Slashdot is in desperate need of is sane moderation.
--fyngyrz
Anon due to mod points. Soylent does it better.
if these pundits are so scary smart, why don't they create the, "3 Laws Safe" program? anybody can critique.
Neither does the devil
Musk and Hawking are frequently going on about branches of science where they have no understanding nor expertise. Why should computers be any different?
Evolution has been "throwing mud at the wall" with complex neural systems for millions of years, and yet we have only the only positive result for sapience. I wouldn't worry about it happening by accident.
Also worth noting: we have zero evidence that disembodied intelligence is even possible - hard to have self-awareness without a self. Something like a self-driving car, with visual processing logic, a 3D model of the world centered on itself, and the need to model/predict actions before taking them - that looks like the human brain, our only example of how to make general intelligence happen. Based on the only actual evidence we have to work-with, that's the sort of platform from which machine intelligence could emerge, not the far wider set of "commercial AI". And that's an area where there's already regulatory oversight and lots of thought about just how autonomous the software should be.
Socialism: a lie told by totalitarians and believed by fools.
When a nascent field such as this is first formalized, people like Feynman and Hawking are invited to join the inaugural faculty, because they bring an elite form of common sense to the party, which helps to pack some real oomph into "formal" as it first takes shape.
This remark is at best a form of ad hominem, once removed, starting from an especially high prominence.
I happen not to agree with Hawking on this issue. All the better. That, too, is how "formal" finally squeezes through its narrow birth canal.
>If both sides are equally strong they will bring in humans to gain the advantage.
Or even if the humans are just more cost-effective. Robots are expensive after all, humans replicate themselves for free and all you have to do is train them.
--- Most topics have many sides worth arguing, allow me to take one opposite you.
Evolution is slow, but a million years (or a hundred million, if we include simpler neural networks) is a long time. Intelligence is incredibly pro-survival, so I'm not sure what your point was there.
Socialism: a lie told by totalitarians and believed by fools.
Thanks for the short story !
At some point, in principle, you can setup regulations, and treat AI research beyond a specific level as precisely equal (or worse than) manufacturing a nuclear weapon.
It would need very concerted global action - or actual willingness to invade and change governments though.
Humans want to do things to other humans.
Even if that is very reprehensible - torturing humans, forcing them to bear children, raping them, subjugating them - requires you to have other humans, and to value their continued existence as a group.
There is no especial reason that AI would do this.
A quick look at the biomass of multi-cellular animals will show you how pro-survival intelligence is: humans are outmassed by our food animals, but that's about it.
If the environment changes for, say, a bear, all it can to is generate random bears in hopes one will be better adapted. Intelligence allows adaption within a lifetime, which is incredibly pro-survival. Also, the simple ability to mentally model an action and see if it kills you before trying it is very pro-survival.
Socialism: a lie told by totalitarians and believed by fools.
and einstein rofls ...
... excuse me sir, but ... what exactly IS formal training in the field here? some 3d spreadsheet pushing data into a loop going over it one by one until it finds a few that work ?
zomg
i dont think they're talking about lvl 4 or 5 autonomous cars your doomsayers over there
Free speech was meant to be free for all... how can anyone grow up in a nanny state ?