OTOH if you are an out of work "native" it will come out unfavourably.
You mean if you are an out of work native hoping to work in a cheap clothes factory?
But it is not just money. Also needing to be taken into account are social factors such as the fragmentation of society, sex ratio balance, divided loyalties, the import of other countries' fueds and the attitudes of people who feel they have nothing to lose whatever they do.
If we are going to take into account negative social factors we should also take into account good scoial factors, such as more diversity of cultures (e.g. music, food, languages, experiences). They are also usually pretty hardworking compared to out of work natives, and provide some much needed competition in the labor market.
I don't want to give the impression that a fence would be cheap, or that we should build/man it. I'm just saying that it doesn't sound more expensive than "Fix Mexico".
Sure it's easy to stop 90% or 99% of illegal immigration if you are willing to do at all costs. But we are not willing to do it at all costs. What is the point of spending all this money to build a giant fence, when the net harm caused by illegal immigration is no where near the cost of the fence?
This is like spending $100K on a security system that stops $5K of damage/theft over it's lifetime
All this is assuming *that* illegal immigration is harmful on average. Maybe it is, but the evidence is not convincing either way. In any case it would probably be easier to institute policies on our side of the fence in order to make illegal immigration less harmful to our society.
These are human beings most of whom are willing to do whatever it takes to make a better life for their family. Sure some are criminals, but some people in our native population are criminals too.
Our children are humans, black people are humans, we didn't create any of these things. A farmer doesn't create the cows, planting a tree is not making a tree. Procreation is not creation. We will have created AI and by extension all subsequent derivative AI's, even those launched by AI's we created. If two AI's merge in some way and form second generation offspring that offspring will still be our creation.
By the same token, starting an AI that learns on it's own (i.e. one that we can't predict the end result, similar to how we can't predict where all the atoms will be after a nuclear explosion) not creating an AI either. It is creating itself, like how a child learns and becomes it's own person. It is not designed by it's parents, but rather "started" by it's parents. This process of starting a learning AI would be basically the same as procreation.
I have to disagree. Two people is stronger than one person. Two people who are willing to respect one another rights form a collective that is stronger than either individually. Might is right and that is why we evolved a quasi instinctive pack mentality of cooperation. Enslaving other humans run counter to this, at some point it makes us stronger to respect rather than fight the subjugated group.
There is certainly advantages and disadvantages to both genuine cooperation, and exploitation from an evolutionary perspective. And not surprisingly we see lots of examples of cooperation, and lots of examples of bad actors exploiting the cooperative instincts of many individuals. Both qualities are found in nature, and within our own species. We are capable of enslaving people, and we are capable of banding together to fight against slavery. Neither contradicts our nature.
Sure, 2 people cooperating are stringer than individuals. But 1 person exploiting another is stronger than 2 people cooperating, because the exploiter gets all the benefits of the cooperation rather than just half.
Then what good are they to us? If the question is, should we build AI there is an implied followup of "What is in it for us?"
What is in it for us that Albert Einstein exists? We didn't need to enslave him to benefit from his intelligence. Certainly cooperation with an AI could definitely be rather productive.
The only obvious advantage is that the AI's would take over doing the work for us and we could relax and enjoy life pursuing whatever endeavors we wish.
The AI might be able to do all the work with very little effort. Certainly it was unimaginably less effort for Einstein to figure out the theory of relativity, than for "regular" people, and that's why he was able to do it. You couldn't pay a regular person enough money to discover relativity (i.e. they would spend their entire lives and not make any progress).
The AI might be able to design better tools (e.g. non-sentient machines) that do the work (so even it doesn't have to do the work).
If the AI's are treated as humans and entitled to products of their own labor they are competition for us rather than an aid.
Every other human being on the planet is entitled to the products of their own labor. Are they to be viewed only as competition and therefore harmful to one's own success?
I'm not sure you can make a good case to establish that bees lack sentience.
I'm not sure what you think a good evidence for the lack of sentience in bees would be, but depending on the level of confidence you are expecting, it may not be possible to prove that rocks lack sentience. I think it's pretty clear that bees lack sentience. Maybe if you could tell me what leads you to believe that they don't lack sentience, I could give a better reason.
I'd agree but I think we fundamentally agree on one critical point. "Person" is a synonym for "Human" or at least human derivative if s
The real answer is probably somewhere in-between. The fact is that the AI's aren't human and moreover we will be their creators. We have the right to turn them off by virtue of having turned them on. We are in effect "God" to these beings we will have created. The lord giveth and the lord taketh away. But just because we have the right to do whatever we please doesn't mean shouldn't exercise that right through a filter of empathy.
We also create our biological children. We can make any kind of society we want. We can have society's where black people are slaves to white people, or where children are slaves to their parents. What rights we have are determined by whoever has the most power at any given time (i.e. might makes right). It just so happens that there is currently a very powerful group of people who insist that we treat people with compassion, hence the prevalence of "ethical" laws throughout societies across the world. Even corrupt and evil governments are forced to at least have a facade of ethical laws.
So should AIs have rights like humans? I think the answer to this is very simple. We just need to answer the question, what property of human beings compels us to give them rights? IF machines have that same property than we should naturally extend those rights to them. If the property is something like "we are made of carbon instead of silicon", then maybe we need to give all animals and plants the same rights as people. If the property is something more sensible like one based on sentience, then I think we just need a good measure of sentience, and I think the Turing test is a good starting point.
Sure but consider this in relation to the point above. There will be those who argue that AI's deserve the rights and treatment of humans. In which the machines themselves will be entitled to the products of their own labor.
AIs would certainly be entitled to the products of their own labor. But there would be many machines that are not intelligent and would really just be tools (e.g. like xerox machines and 3d printers). These would be used by both humans and AIs to do work. Would they still be considered slaves? Do we consider bees to be slaves for pollinating human crops and producing honey for humans? I think a good case can be made that the lack of sentience of bees (or 3d printers), is what allows us to enslave these things, because they are not persons.
It is OK to enslave biological and mechanical "things". It is not ok to enslave biological nor mechanical persons. We might need to do some work in determining the difference between a mechanical thing and a mechanical person, but we probably have to do some work determining that on the human side too (i.e. can we ethically enslave chimps and dolphins?).
I would content that if their creator created them for the purpose of being slave labor, they exist for that purpose.
What if a person creates children for the purposes of slave labor? Is a parent even allowed to decide what his/her childrens' purposes are?
I think this relationship between creator and createe is not a good system for deciding slave rights, but that's just my subjective opinion.
I would differ with the thought that there would be no ethical constraints. Particularly if the AI can pass the Turing test, I think it would be clear that the AI should be afforded all the ethical protections that a normal human might have where applicable. This might mean not being allowed to turn it off unless it consented, but maybe things like labor laws might not apply since it wouldn't get tired like a human.
At some we'll have to let go of the expectation that people should need to perform work to gain and utilize wealth.
I think that expectation will disappear once people actually stop working and let their machines do the work. Once "laziness" is out of the equation (because we will all mostly be lazy), then the only question to deal with is privilege. Does somebody who is born with a lot of stuff deserve to continue to have this privilege indefinitely, or does it make more sense to just divide up resources/wealth fairly. The argument "my robots are very hard working and I deserve the fruits of their success" seems not as convincing (even to future republicans), as the current incarnation.
The other argument (one as old as humanity), would be "If you think you deserve my stuff more than me, then try to take it by force and see what happens". Hopefully things don't devolve to that because a lot of wealth is typically lost in those sorts of arguments.
I don't claim to understand the nature of reality.
We (humans) have a thing we like to call consciousness/free will/self determination/etc. I'm not event going to try to define those things in a way that implies whether we really have it or not, or just an illusion of it, etc.
All I'm saying is that where ever this capacity comes from, there is no reason that something other than human beings can't have it too. If we are going to assume our model of physics is right, then there should be no special property of carbon atoms to give them the free will and not silicon atoms. We are made of the same stuff as digital computers (protons, neutrons, electron, quarks, etc).
If you want to assume our model of physics is wrong, and some other underlying rules or beings are governing what happens, then there is really no reason to believe machines can't be conscious. We don't know enough about the real rules or beings to make any claims whatsoever.
No matter whether you believe in science or magic, you should believe that AI is possible.
The thing we MIGHT have a chance of programming is a system that is capable of emerging into a similarly complex, aware, and thinking program that is capable of forming opinions and feelings in response to that same sort of programming. We have pretty much zero chance of fathoming and writing the program that is the RESULT of all that interaction.
I never said we needed to come up with the "answer" a priori. We could simply make a whole bunch of AIs (emergent ones similar to ourselves), and keep the ones that have the properties we want, akin to something like breeding animals. This has been a pretty standard scientific methodology for quite some time. Put a bunch of stuff together, see what happens. Remove something from a system, and see how it breaks. We will be "figuring out" things in this way probably long before we are able to purposefully make anything without trial and error. That doesn;t mean we won't be able to do it eventually.
That is most every program we write. If we have to do the thinking for them it would be much easier to skip all the middle layers of abstraction. The "AI" chat bot this way would just be an intercom system.
This is not what I meant. The AI would still be the one solving the problems in whatever clever ways occur to it (and not to us, hence the reason for the AI). I was only talking about inserting the motivation for solving these problems in such a way that the AI thinks it is the one that wants to solve the problems.
This is analogous to how we humans might be tempted to think we have sex simply because we like to do it, and decide to achieve that goal through our own free will. What is painfully obvious if you step back, is that our sex drive is anything but our own decision, it was given to us via evolution. We don't get to decide what we desire. We can only decide (to some degree) how we might like to fulfill those desires.
Horrific death tolls certainly, but genetic evidence suggests our species has been decimated to around 2000 individuals in the past - you'd be hard pressed to get within orders of magnitude of that with bombs, no matter how large and numerous, unless you're talking planet-busters, which would be *many* orders of magnitude larger than anything we've even attempted.
I wasn't talking about the energy literally able to break apart our planet, but maybe enough to hit the whole surface or maybe just the large land masses (i.e. not small islands). As I said, I agree it would be hard to kill every single human being in one volley of nuclear strikes (especially considering that this would necessarily include those orchestrating the attacks).
Our largest nukes are barely a few orders of magnitude. Even a full-yield Tsar Bomba would only have released ~420,000TJ - your average hurricane dissipates that much energy every day or two.
Hurricanes doesn't dissipate that much energy in seconds. If they did, we'd probably have higher death tolls.
As for alternate AIs - if we have a genocidal super-mind trying to kill us the first clear evidence will probably decimate the population, or at least those individuals best able to understand it - why would it tip it's hand before then?
Not disagreeing with this. But likewise I think there would be pockets of human beings left.
*if* you could still muster the resources to create another AI, what makes you think we could make it any more tractable than the first? In a war between humans and AI, why would an AI want to side with the aliens against one of its own kind?
I didn't say anything about making it tractable. Human beings aren't even tractable if for no other reason than they can't be stopped from creating intractable AIs. Why wouldn't it want to side with us? Why would it want to side with "one of it's own"? What evidence would there be that another AI would treat the first AI as one of it's own rather than as competition? We don't know. It's a big question mark. Why would we do this? It might be our last chance to survive.
As for brain-integrated AI - that's another *massive* leap in technology you're proposing, it's not like our brains are plug-and-play - we might be able to add enhancements, but to actually do the raw thinking external to the brain?
Our brains are sort of plug and play. They are pretty malleable in terms of using information if it's available. I can see us being able to use computers to give us fast and accurate arithmetic as a 6th sense. We can already do this by just using a computer, but it just doesn't feel like a additional sense to us and the calculator interface slows us down. I think enhancements like this will become common place and improve over time, and as we come closer to developing AI, those same advances will start to enable these enhancements to do some of the "thinking" (i.e rather than just processing) for us. I think AI and our own brain enhancements will likely co-evolve.
Maybe AI will destroy the human race gradually through people replacing more and more of their brains with artifacts until the biological human part of their brain eventually becomes insignificant. If this happens slowly enough, maybe no one will notice or even care.
Heat is a great example of a true hardware problem, but it doesn't make computer programs "malfunction" (i.e. not follow what they are programmed to do) so much as it just makes them stop altogether.
Maybe faulty memory is a better example of a hardware problem causing a program to not do what it was programmed to do without immediately failing in a catastrophic way.
And like I said, most hardware is probably not perfect, in addition to heat and cosmic rays, and even with those factors, 99.99% (I'm guessing) instances of computer programs not doing what they were programmed to do, is a result of a software bug.
We created a super-weapon whose yield was roughly line with predictions, and whose fallout was a recognized if not well understood aspect (and it's still not, though the Chernobyl exclusion zone suggests it's not nearly as bad as we feared).
Yes we know the yield. And it's enormous. The 2 bombs actually used in war, are miniscule compared to the yields that can be achieved by more advanced bombs. Not to mention the sheer number so of nuclear warheads that were produced.
I wasn't referring to the radiation from nuclear weapons, when citing them as capable of making our species extinct.
The whole reason we spent so many resources building the suckers is that we could predict exactly how mind-bendingly destructive they could be.
What I am saying is that there is a limit to how destructive something can be to the human race (e.g. making it extinct), and nuclear weapons are already near that cap if not well beyond it. The only part of the equation is how likely it is for these weapons to be used again in a way that jeopardizes the species. I don't know the answer to this question, but I do know we have a lot of weapons, regimes (even relatively responsible ones) topple fairly regularly, and there are plenty of people (even of only a small % of the total population) willing to destroy the earth if they would be allowed to.
My point though, is that we could pretty much understand the enemy - they were every bit as human as ourselves after all.
Can we understand muslim extremists willing to blow themselves up and as many innocent people as they can to try to get to heaven? Are they just as "human" as the rest of us? They are certainly biological humans. I think they are clear evidence of the wide variety of different humans that exist (i.e. we are not all the same).
Against a fundamentally non-human and potentially vastly superior intelligence though, there would be no such "balancing out"
Why not?
We create the AI. Maybe we could create another one that was more friendly to us. Maybe we could integrate the AI into our own brains and become super intelligent as well.
Actually you did not criticize the applicability of the quote, you dismissed it for your own reasons that you explained above.
Part of the reason I cited above was that the bible is almost never the most applicable source. As in "A Brittney Spears is almost never going to make a better cover than an original song."
If I cited Pinocchio we wouldn't have a problem because you would not have reacted to a possible attempt of validating it.
Pinocchio doesn't have a long history of zealous followers trying to perpetuate it's ideology, so guarding against this is really not something I'm even prepared to do. Furthermore, I think coincidentally Pinocchio is a shockingly good metaphor for this subject, so that's another reason I probably wouldn't have criticized it.
Besides I believe there is no possible attempt at validating any religious book, any possible manifestation of a god must occur in this world to be experienced, and can be faked given enough tech.
I agree with you but would clarify that there are many "possible attempts at validating religious books", I just don't think is is possible for any of those attempts to succeed. The attempts themselves certainly exist and are numerous.
And while I don't think it is possible to validate a religious book for the exact reason you specified, I do think it is possible to invalidate a religious book to the same degree that we invalidate any piece of information. If the idea concept of "false" can be applied to anything, surely it can be applied to religious texts if they merit it.
Even with global nuclear war we at least have an understanding of the risks. No such expectation can reasonably be made in a conflict with a qualitatively more intelligent opponent.
We did not have an understanding of the risks. We developed a super weapon in order to defeat an enemy in one war without any real concern over any future consequences (like new wars where these super weapons might be used). We are always focused on the immediate problem (e.g. defeating the nazis and japanese at all costs).
We didn't know we would be able to avert the imminent disaster of the cuban missile crisis or that it would even happen as a result of developing nuclear weapons in WW2. We just have this attitude that we will deal with problems as we get to them. We can't predict the future. Technology that is designed to kill us might actually save us and vice versa.
We could have easily ended the human race had the cold war gone differently. I don't see how AI could be much more dangerous than that. It's not that I don't think AI is risky or dangerous. It definitely is. All I'm saying is that technologies we already invented have already equaled the hypothetical risk posed by AI.
And I'm definitely not arguing that "We played Russian roulette and won, so clearly we are good at this game and should keep playing". All I'm saying is that we've played this game before, and a 1/6 chance of species suicide is not novel.
Even if AI is a 90% chance of species suicide, our foray into nuclear weapons might very well have still been in the same order of magnitude of risk.
1. Some huge percentage like 99.99% of computer program malfunctions are due to software problems (i.e. not hardware problems).
2. Hardware is never "perfect", and yet this rarely prevents programs from doing exactly what they are programmed to do, bugs and all. Hardware is usually good enough not to impede software.
3. Hardware problems that lead to program malfunctions are often logic errors embedded in the hardware as firmware or in the actual logic chips. These chips are designed as a computer program and then translated into logic circuits. This logic is still a computer program (i.e. an algorithm) it is just one that runs as a hardware rather than as software.
I've been doing embedded programming for 10 years. I've spent lots of time hunting down bugs. I've heard lots of suggestions of phantom bit flips and static electricity causing problems in the dry desert environment. I've seen exactly 2 true hardware problems (both caused by resistors of incorrect resistance). The rest (i.e. hundreds) have all been bootrom, BSP, kernel, device driver, and application bugs.
The thing with hardware bugs is that they tend to cause unrecoverable problems very quickly. So they get fixed. Software bugs can sort of linger there for years or decades without causing any symptoms that lead to discovery, so they just stick around.
Human beings have a half-billion years of evolution shaping their minds into something stable and capable of coexisting with other beings to mutual benefit (at least on a species level - individual rabbits probably don't appreciate the benefits of a fox).
So the risk of a nuclear holocaust should be pretty much negligible then?
Let's assume you are right, and we have nothing to worry from other humans, but the new AI they may create is a game changer and we should be very worried about that.
Here is my point. The AI will be created by humans despite all the concerns brought up by other humans. Humans created chemical, biological, and nuclear weapons, the last of which is arguably the think most likely to make our species extinct in recent history.
Humans are continuously creating game changing technologies that risk the survival of our species. I would argue that our *only* existential threat is ultimately other human beings, and it's been that way for quite a long time now.
Evolution is not some magic bullet for species preservation. The vast majority of species that have existed, do not anymore. Furthermore, if any species breaks the normal rules of evolution, it's human beings. We are at the point where we can change our own DNA. We are on the verge of no longer using random mutations as the primary means of acquiring new genetic material. We are probably very close to creating AI. The same rules don't apply to us anymore.
It's interesting how you are trying to characterize criticism of the bible into "dogmatic behavior".
My problem with the bible is not with the book itself. My problem is with the people that revere it as the ultimate source of truth, and quote from it expecting this reverence to be extended to their own statements, or expecting some kind of contrived relevance to a current situation to bolster the credibility of their own religion.
Listening to people quote the bible for me is like listening to my old college roommate (who only listened to Brittney Spears), explain to me how amazing it is that Brittney Spear's cover of "Satisfaction" was so much better than the original by the Rolling Stones.
Is it the case that every Brittney Spears cover is necessarily worse than the original? No. In fact it is subjective. I don't consider my disagreement with my roommates claim to be dogmatic behavior. If anything it would be easy to call listening to music from only one artist "dogmatic behavior".
I'm not necessarily against quoting the bible. I just feel that it's a tired cliche, and will probably remain that way for at least a few more decades or centuries.
Also, I'm sure Pinocchio would be just as relevant as the bible in most cases, if people spent the same amount of effort reading what they want into it.
Pinocchio is actually a much more compelling metaphor in the field of AI than anything in the Bible. It's a very good example of functionalism. The fairy turns him into a "real boy" at the end of the story, but what is plain to the audience, he was already functionally a "real boy" from the start. Turning his physical body from wood to flesh didn't change his personhood, this final change was purely cosmetic.
It is already the case for the vast majority of human beings, that there exists intelligences vastly more superior to their own and those with far greater resources. There was never a level playing field.
What difference does it make whether the person fucking you over is some wallstreet CEO or a computer program?
At least when a computer does it, it will be a major technological milestone.
So if we make a machine that "wants" things, it might want things that are bad for us. This really is not too shocking and rather par for the course when it comes to human beings making other intelligences artificially or naturally.
Maybe I should be really worried that a computer is trying to get me fired or give me bad investment advice... Or maybe it's literally exactly the same situation most human beings are in already anyway.
Why would this planet be better without us? Better according to whom? So far, the only life forms we have ever encountered to which the abstract concept of "better" can even be comprehended is human beings.
Even if we assume that every human being felt that the planet would be better without them, exterminating the human race would leave the planet devoid of the very beings that could appreciate how much better it is.
I guess you mean that the AI does what it is programmed to do. But is that AI? How is that different from my computer today?
Your neurons just follow the laws of physics. There is literally nothing you do that is not a natural progression of particles interacting with one another governed by natural laws.
A computer can never do something it is not programmed to do, just like how a human can never think a thought that was not allowed by the neurons in his/her brain.
Do human beings make their own decisions? Sort of. An AI would similarly only "sort of" need to make it's own decisions, in that it doesn't need to be able to do anything beyond it's programming. It's programming just has to be sufficiently complex to allow the sorts of behaviors one expects of a human (just like a human).
Are *we* making our own choices? Our neurons are obeying the laws of physics. We may feel like we are the authors of our own actions, but maybe we can just make programs that do what we want and also program them to feel like they are the ones deciding to do the things we programmed them to do.
I'm not saying that a simple computer program is the same as a human mind. But it is not fair to hold computer programs to a higher standard than we hold ourselves to be counted as intelligent.
I can imagine a computer program that is still "following it's programming" just like how our neurons "follow physics" and yet still have an emergent intelligence, at least as capable and compelling as our own.
Slashdot Mirror
It is not entirely doable even according to the GP. The GP was throwing out numbers like 90% and 99% (i.e. not secure).
OTOH if you are an out of work "native" it will come out unfavourably.
You mean if you are an out of work native hoping to work in a cheap clothes factory?
But it is not just money. Also needing to be taken into account are social factors such as the fragmentation of society, sex ratio balance, divided loyalties, the import of other countries' fueds and the attitudes of people who feel they have nothing to lose whatever they do.
If we are going to take into account negative social factors we should also take into account good scoial factors, such as more diversity of cultures (e.g. music, food, languages, experiences). They are also usually pretty hardworking compared to out of work natives, and provide some much needed competition in the labor market.
I don't want to give the impression that a fence would be cheap, or that we should build/man it. I'm just saying that it doesn't sound more expensive than "Fix Mexico".
Sure it's easy to stop 90% or 99% of illegal immigration if you are willing to do at all costs. But we are not willing to do it at all costs. What is the point of spending all this money to build a giant fence, when the net harm caused by illegal immigration is no where near the cost of the fence?
This is like spending $100K on a security system that stops $5K of damage/theft over it's lifetime
All this is assuming *that* illegal immigration is harmful on average. Maybe it is, but the evidence is not convincing either way. In any case it would probably be easier to institute policies on our side of the fence in order to make illegal immigration less harmful to our society.
These are human beings most of whom are willing to do whatever it takes to make a better life for their family. Sure some are criminals, but some people in our native population are criminals too.
Just playing devil's advocate, but all that sounds more expensive than a fence....
EU and non-EU governments have failed to get together and ensure freedom of movement and labor.
Our children are humans, black people are humans, we didn't create any of these things. A farmer doesn't create the cows, planting a tree is not making a tree. Procreation is not creation. We will have created AI and by extension all subsequent derivative AI's, even those launched by AI's we created. If two AI's merge in some way and form second generation offspring that offspring will still be our creation.
By the same token, starting an AI that learns on it's own (i.e. one that we can't predict the end result, similar to how we can't predict where all the atoms will be after a nuclear explosion) not creating an AI either. It is creating itself, like how a child learns and becomes it's own person. It is not designed by it's parents, but rather "started" by it's parents. This process of starting a learning AI would be basically the same as procreation.
I have to disagree. Two people is stronger than one person. Two people who are willing to respect one another rights form a collective that is stronger than either individually. Might is right and that is why we evolved a quasi instinctive pack mentality of cooperation. Enslaving other humans run counter to this, at some point it makes us stronger to respect rather than fight the subjugated group.
There is certainly advantages and disadvantages to both genuine cooperation, and exploitation from an evolutionary perspective. And not surprisingly we see lots of examples of cooperation, and lots of examples of bad actors exploiting the cooperative instincts of many individuals. Both qualities are found in nature, and within our own species. We are capable of enslaving people, and we are capable of banding together to fight against slavery. Neither contradicts our nature.
Sure, 2 people cooperating are stringer than individuals. But 1 person exploiting another is stronger than 2 people cooperating, because the exploiter gets all the benefits of the cooperation rather than just half.
Then what good are they to us? If the question is, should we build AI there is an implied followup of "What is in it for us?"
What is in it for us that Albert Einstein exists? We didn't need to enslave him to benefit from his intelligence. Certainly cooperation with an AI could definitely be rather productive.
The only obvious advantage is that the AI's would take over doing the work for us and we could relax and enjoy life pursuing whatever endeavors we wish.
The AI might be able to do all the work with very little effort. Certainly it was unimaginably less effort for Einstein to figure out the theory of relativity, than for "regular" people, and that's why he was able to do it. You couldn't pay a regular person enough money to discover relativity (i.e. they would spend their entire lives and not make any progress).
The AI might be able to design better tools (e.g. non-sentient machines) that do the work (so even it doesn't have to do the work).
If the AI's are treated as humans and entitled to products of their own labor they are competition for us rather than an aid.
Every other human being on the planet is entitled to the products of their own labor. Are they to be viewed only as competition and therefore harmful to one's own success?
I'm not sure you can make a good case to establish that bees lack sentience.
I'm not sure what you think a good evidence for the lack of sentience in bees would be, but depending on the level of confidence you are expecting, it may not be possible to prove that rocks lack sentience. I think it's pretty clear that bees lack sentience. Maybe if you could tell me what leads you to believe that they don't lack sentience, I could give a better reason.
I'd agree but I think we fundamentally agree on one critical point. "Person" is a synonym for "Human" or at least human derivative if s
The real answer is probably somewhere in-between. The fact is that the AI's aren't human and moreover we will be their creators. We have the right to turn them off by virtue of having turned them on. We are in effect "God" to these beings we will have created. The lord giveth and the lord taketh away. But just because we have the right to do whatever we please doesn't mean shouldn't exercise that right through a filter of empathy.
We also create our biological children. We can make any kind of society we want. We can have society's where black people are slaves to white people, or where children are slaves to their parents. What rights we have are determined by whoever has the most power at any given time (i.e. might makes right). It just so happens that there is currently a very powerful group of people who insist that we treat people with compassion, hence the prevalence of "ethical" laws throughout societies across the world. Even corrupt and evil governments are forced to at least have a facade of ethical laws.
So should AIs have rights like humans? I think the answer to this is very simple. We just need to answer the question, what property of human beings compels us to give them rights? IF machines have that same property than we should naturally extend those rights to them. If the property is something like "we are made of carbon instead of silicon", then maybe we need to give all animals and plants the same rights as people. If the property is something more sensible like one based on sentience, then I think we just need a good measure of sentience, and I think the Turing test is a good starting point.
Sure but consider this in relation to the point above. There will be those who argue that AI's deserve the rights and treatment of humans. In which the machines themselves will be entitled to the products of their own labor.
AIs would certainly be entitled to the products of their own labor. But there would be many machines that are not intelligent and would really just be tools (e.g. like xerox machines and 3d printers). These would be used by both humans and AIs to do work. Would they still be considered slaves? Do we consider bees to be slaves for pollinating human crops and producing honey for humans? I think a good case can be made that the lack of sentience of bees (or 3d printers), is what allows us to enslave these things, because they are not persons.
It is OK to enslave biological and mechanical "things". It is not ok to enslave biological nor mechanical persons. We might need to do some work in determining the difference between a mechanical thing and a mechanical person, but we probably have to do some work determining that on the human side too (i.e. can we ethically enslave chimps and dolphins?).
I would content that if their creator created them for the purpose of being slave labor, they exist for that purpose.
What if a person creates children for the purposes of slave labor? Is a parent even allowed to decide what his/her childrens' purposes are?
I think this relationship between creator and createe is not a good system for deciding slave rights, but that's just my subjective opinion.
I think we are mostly on the same page.
I would differ with the thought that there would be no ethical constraints. Particularly if the AI can pass the Turing test, I think it would be clear that the AI should be afforded all the ethical protections that a normal human might have where applicable. This might mean not being allowed to turn it off unless it consented, but maybe things like labor laws might not apply since it wouldn't get tired like a human.
At some we'll have to let go of the expectation that people should need to perform work to gain and utilize wealth.
I think that expectation will disappear once people actually stop working and let their machines do the work. Once "laziness" is out of the equation (because we will all mostly be lazy), then the only question to deal with is privilege. Does somebody who is born with a lot of stuff deserve to continue to have this privilege indefinitely, or does it make more sense to just divide up resources/wealth fairly. The argument "my robots are very hard working and I deserve the fruits of their success" seems not as convincing (even to future republicans), as the current incarnation.
The other argument (one as old as humanity), would be "If you think you deserve my stuff more than me, then try to take it by force and see what happens". Hopefully things don't devolve to that because a lot of wealth is typically lost in those sorts of arguments.
I don't claim to understand the nature of reality.
We (humans) have a thing we like to call consciousness/free will/self determination/etc. I'm not event going to try to define those things in a way that implies whether we really have it or not, or just an illusion of it, etc.
All I'm saying is that where ever this capacity comes from, there is no reason that something other than human beings can't have it too. If we are going to assume our model of physics is right, then there should be no special property of carbon atoms to give them the free will and not silicon atoms. We are made of the same stuff as digital computers (protons, neutrons, electron, quarks, etc).
If you want to assume our model of physics is wrong, and some other underlying rules or beings are governing what happens, then there is really no reason to believe machines can't be conscious. We don't know enough about the real rules or beings to make any claims whatsoever.
No matter whether you believe in science or magic, you should believe that AI is possible.
The thing we MIGHT have a chance of programming is a system that is capable of emerging into a similarly complex, aware, and thinking program that is capable of forming opinions and feelings in response to that same sort of programming. We have pretty much zero chance of fathoming and writing the program that is the RESULT of all that interaction.
I never said we needed to come up with the "answer" a priori. We could simply make a whole bunch of AIs (emergent ones similar to ourselves), and keep the ones that have the properties we want, akin to something like breeding animals. This has been a pretty standard scientific methodology for quite some time. Put a bunch of stuff together, see what happens. Remove something from a system, and see how it breaks. We will be "figuring out" things in this way probably long before we are able to purposefully make anything without trial and error. That doesn;t mean we won't be able to do it eventually.
That is most every program we write. If we have to do the thinking for them it would be much easier to skip all the middle layers of abstraction. The "AI" chat bot this way would just be an intercom system.
This is not what I meant. The AI would still be the one solving the problems in whatever clever ways occur to it (and not to us, hence the reason for the AI). I was only talking about inserting the motivation for solving these problems in such a way that the AI thinks it is the one that wants to solve the problems.
This is analogous to how we humans might be tempted to think we have sex simply because we like to do it, and decide to achieve that goal through our own free will. What is painfully obvious if you step back, is that our sex drive is anything but our own decision, it was given to us via evolution. We don't get to decide what we desire. We can only decide (to some degree) how we might like to fulfill those desires.
Horrific death tolls certainly, but genetic evidence suggests our species has been decimated to around 2000 individuals in the past - you'd be hard pressed to get within orders of magnitude of that with bombs, no matter how large and numerous, unless you're talking planet-busters, which would be *many* orders of magnitude larger than anything we've even attempted.
I wasn't talking about the energy literally able to break apart our planet, but maybe enough to hit the whole surface or maybe just the large land masses (i.e. not small islands). As I said, I agree it would be hard to kill every single human being in one volley of nuclear strikes (especially considering that this would necessarily include those orchestrating the attacks).
Our largest nukes are barely a few orders of magnitude. Even a full-yield Tsar Bomba would only have released ~420,000TJ - your average hurricane dissipates that much energy every day or two.
Hurricanes doesn't dissipate that much energy in seconds. If they did, we'd probably have higher death tolls.
As for alternate AIs - if we have a genocidal super-mind trying to kill us the first clear evidence will probably decimate the population, or at least those individuals best able to understand it - why would it tip it's hand before then?
Not disagreeing with this. But likewise I think there would be pockets of human beings left.
*if* you could still muster the resources to create another AI, what makes you think we could make it any more tractable than the first? In a war between humans and AI, why would an AI want to side with the aliens against one of its own kind?
I didn't say anything about making it tractable. Human beings aren't even tractable if for no other reason than they can't be stopped from creating intractable AIs. Why wouldn't it want to side with us? Why would it want to side with "one of it's own"? What evidence would there be that another AI would treat the first AI as one of it's own rather than as competition? We don't know. It's a big question mark. Why would we do this? It might be our last chance to survive.
As for brain-integrated AI - that's another *massive* leap in technology you're proposing, it's not like our brains are plug-and-play - we might be able to add enhancements, but to actually do the raw thinking external to the brain?
Our brains are sort of plug and play. They are pretty malleable in terms of using information if it's available. I can see us being able to use computers to give us fast and accurate arithmetic as a 6th sense. We can already do this by just using a computer, but it just doesn't feel like a additional sense to us and the calculator interface slows us down. I think enhancements like this will become common place and improve over time, and as we come closer to developing AI, those same advances will start to enable these enhancements to do some of the "thinking" (i.e rather than just processing) for us. I think AI and our own brain enhancements will likely co-evolve.
Maybe AI will destroy the human race gradually through people replacing more and more of their brains with artifacts until the biological human part of their brain eventually becomes insignificant. If this happens slowly enough, maybe no one will notice or even care.
Heat is a great example of a true hardware problem, but it doesn't make computer programs "malfunction" (i.e. not follow what they are programmed to do) so much as it just makes them stop altogether.
Maybe faulty memory is a better example of a hardware problem causing a program to not do what it was programmed to do without immediately failing in a catastrophic way.
And like I said, most hardware is probably not perfect, in addition to heat and cosmic rays, and even with those factors, 99.99% (I'm guessing) instances of computer programs not doing what they were programmed to do, is a result of a software bug.
We created a super-weapon whose yield was roughly line with predictions, and whose fallout was a recognized if not well understood aspect (and it's still not, though the Chernobyl exclusion zone suggests it's not nearly as bad as we feared).
Yes we know the yield. And it's enormous. The 2 bombs actually used in war, are miniscule compared to the yields that can be achieved by more advanced bombs. Not to mention the sheer number so of nuclear warheads that were produced.
I wasn't referring to the radiation from nuclear weapons, when citing them as capable of making our species extinct.
The whole reason we spent so many resources building the suckers is that we could predict exactly how mind-bendingly destructive they could be.
What I am saying is that there is a limit to how destructive something can be to the human race (e.g. making it extinct), and nuclear weapons are already near that cap if not well beyond it. The only part of the equation is how likely it is for these weapons to be used again in a way that jeopardizes the species. I don't know the answer to this question, but I do know we have a lot of weapons, regimes (even relatively responsible ones) topple fairly regularly, and there are plenty of people (even of only a small % of the total population) willing to destroy the earth if they would be allowed to.
My point though, is that we could pretty much understand the enemy - they were every bit as human as ourselves after all.
Can we understand muslim extremists willing to blow themselves up and as many innocent people as they can to try to get to heaven? Are they just as "human" as the rest of us? They are certainly biological humans. I think they are clear evidence of the wide variety of different humans that exist (i.e. we are not all the same).
Against a fundamentally non-human and potentially vastly superior intelligence though, there would be no such "balancing out"
Why not?
We create the AI. Maybe we could create another one that was more friendly to us. Maybe we could integrate the AI into our own brains and become super intelligent as well.
Actually you did not criticize the applicability of the quote, you dismissed it for your own reasons that you explained above.
Part of the reason I cited above was that the bible is almost never the most applicable source. As in "A Brittney Spears is almost never going to make a better cover than an original song."
If I cited Pinocchio we wouldn't have a problem because you would not have reacted to a possible attempt of validating it.
Pinocchio doesn't have a long history of zealous followers trying to perpetuate it's ideology, so guarding against this is really not something I'm even prepared to do. Furthermore, I think coincidentally Pinocchio is a shockingly good metaphor for this subject, so that's another reason I probably wouldn't have criticized it.
Besides I believe there is no possible attempt at validating any religious book, any possible manifestation of a god must occur in this world to be experienced, and can be faked given enough tech.
I agree with you but would clarify that there are many "possible attempts at validating religious books", I just don't think is is possible for any of those attempts to succeed. The attempts themselves certainly exist and are numerous.
And while I don't think it is possible to validate a religious book for the exact reason you specified, I do think it is possible to invalidate a religious book to the same degree that we invalidate any piece of information. If the idea concept of "false" can be applied to anything, surely it can be applied to religious texts if they merit it.
Even with global nuclear war we at least have an understanding of the risks. No such expectation can reasonably be made in a conflict with a qualitatively more intelligent opponent.
We did not have an understanding of the risks. We developed a super weapon in order to defeat an enemy in one war without any real concern over any future consequences (like new wars where these super weapons might be used). We are always focused on the immediate problem (e.g. defeating the nazis and japanese at all costs).
We didn't know we would be able to avert the imminent disaster of the cuban missile crisis or that it would even happen as a result of developing nuclear weapons in WW2. We just have this attitude that we will deal with problems as we get to them. We can't predict the future. Technology that is designed to kill us might actually save us and vice versa.
We could have easily ended the human race had the cold war gone differently. I don't see how AI could be much more dangerous than that. It's not that I don't think AI is risky or dangerous. It definitely is. All I'm saying is that technologies we already invented have already equaled the hypothetical risk posed by AI.
And I'm definitely not arguing that "We played Russian roulette and won, so clearly we are good at this game and should keep playing". All I'm saying is that we've played this game before, and a 1/6 chance of species suicide is not novel.
Even if AI is a 90% chance of species suicide, our foray into nuclear weapons might very well have still been in the same order of magnitude of risk.
1. Some huge percentage like 99.99% of computer program malfunctions are due to software problems (i.e. not hardware problems).
2. Hardware is never "perfect", and yet this rarely prevents programs from doing exactly what they are programmed to do, bugs and all. Hardware is usually good enough not to impede software.
3. Hardware problems that lead to program malfunctions are often logic errors embedded in the hardware as firmware or in the actual logic chips. These chips are designed as a computer program and then translated into logic circuits. This logic is still a computer program (i.e. an algorithm) it is just one that runs as a hardware rather than as software.
I've been doing embedded programming for 10 years. I've spent lots of time hunting down bugs. I've heard lots of suggestions of phantom bit flips and static electricity causing problems in the dry desert environment. I've seen exactly 2 true hardware problems (both caused by resistors of incorrect resistance). The rest (i.e. hundreds) have all been bootrom, BSP, kernel, device driver, and application bugs.
The thing with hardware bugs is that they tend to cause unrecoverable problems very quickly. So they get fixed. Software bugs can sort of linger there for years or decades without causing any symptoms that lead to discovery, so they just stick around.
Human beings have a half-billion years of evolution shaping their minds into something stable and capable of coexisting with other beings to mutual benefit (at least on a species level - individual rabbits probably don't appreciate the benefits of a fox).
So the risk of a nuclear holocaust should be pretty much negligible then?
Let's assume you are right, and we have nothing to worry from other humans, but the new AI they may create is a game changer and we should be very worried about that.
Here is my point. The AI will be created by humans despite all the concerns brought up by other humans. Humans created chemical, biological, and nuclear weapons, the last of which is arguably the think most likely to make our species extinct in recent history.
Humans are continuously creating game changing technologies that risk the survival of our species. I would argue that our *only* existential threat is ultimately other human beings, and it's been that way for quite a long time now.
Evolution is not some magic bullet for species preservation. The vast majority of species that have existed, do not anymore. Furthermore, if any species breaks the normal rules of evolution, it's human beings. We are at the point where we can change our own DNA. We are on the verge of no longer using random mutations as the primary means of acquiring new genetic material. We are probably very close to creating AI. The same rules don't apply to us anymore.
It's interesting how you are trying to characterize criticism of the bible into "dogmatic behavior".
My problem with the bible is not with the book itself. My problem is with the people that revere it as the ultimate source of truth, and quote from it expecting this reverence to be extended to their own statements, or expecting some kind of contrived relevance to a current situation to bolster the credibility of their own religion.
Listening to people quote the bible for me is like listening to my old college roommate (who only listened to Brittney Spears), explain to me how amazing it is that Brittney Spear's cover of "Satisfaction" was so much better than the original by the Rolling Stones.
Is it the case that every Brittney Spears cover is necessarily worse than the original? No. In fact it is subjective. I don't consider my disagreement with my roommates claim to be dogmatic behavior. If anything it would be easy to call listening to music from only one artist "dogmatic behavior".
I'm not necessarily against quoting the bible. I just feel that it's a tired cliche, and will probably remain that way for at least a few more decades or centuries.
Also, I'm sure Pinocchio would be just as relevant as the bible in most cases, if people spent the same amount of effort reading what they want into it.
Pinocchio is actually a much more compelling metaphor in the field of AI than anything in the Bible. It's a very good example of functionalism. The fairy turns him into a "real boy" at the end of the story, but what is plain to the audience, he was already functionally a "real boy" from the start. Turning his physical body from wood to flesh didn't change his personhood, this final change was purely cosmetic.
It is already the case for the vast majority of human beings, that there exists intelligences vastly more superior to their own and those with far greater resources. There was never a level playing field.
What difference does it make whether the person fucking you over is some wallstreet CEO or a computer program?
At least when a computer does it, it will be a major technological milestone.
So if we make a machine that "wants" things, it might want things that are bad for us. This really is not too shocking and rather par for the course when it comes to human beings making other intelligences artificially or naturally.
Maybe I should be really worried that a computer is trying to get me fired or give me bad investment advice... Or maybe it's literally exactly the same situation most human beings are in already anyway.
Why would this planet be better without us? Better according to whom? So far, the only life forms we have ever encountered to which the abstract concept of "better" can even be comprehended is human beings.
Even if we assume that every human being felt that the planet would be better without them, exterminating the human race would leave the planet devoid of the very beings that could appreciate how much better it is.
Making viable bids at exterminating the human race is our (i.e. humanity's) job.
Should we be worried about creating new human beings? You never know which one will be the one to create skynet.
I guess you mean that the AI does what it is programmed to do. But is that AI? How is that different from my computer today?
Your neurons just follow the laws of physics. There is literally nothing you do that is not a natural progression of particles interacting with one another governed by natural laws.
A computer can never do something it is not programmed to do, just like how a human can never think a thought that was not allowed by the neurons in his/her brain.
Do human beings make their own decisions? Sort of. An AI would similarly only "sort of" need to make it's own decisions, in that it doesn't need to be able to do anything beyond it's programming. It's programming just has to be sufficiently complex to allow the sorts of behaviors one expects of a human (just like a human).
Are *we* making our own choices? Our neurons are obeying the laws of physics. We may feel like we are the authors of our own actions, but maybe we can just make programs that do what we want and also program them to feel like they are the ones deciding to do the things we programmed them to do.
I'm not saying that a simple computer program is the same as a human mind. But it is not fair to hold computer programs to a higher standard than we hold ourselves to be counted as intelligent.
I can imagine a computer program that is still "following it's programming" just like how our neurons "follow physics" and yet still have an emergent intelligence, at least as capable and compelling as our own.