the last invention we would ever need…
One of the ideas for the moment of the Technological Singularity focuses around the potential invention of an artificial intelligence so flexible and omnipotent, it could pretty much take over all our day to day tasks. It could design new machines to do ever more sophisticated jobs. It could troubleshoot, fix and upgrade itself without the slightest bit of human intervention. It could run our cities and cure dangerous and new diseases. As said by philosopher and futurist Nick Bostrom, this hyper-intelligent machine would be the final invention we would ever need, making any further efforts to do, well, anything on our part, totally unnecessary since the machines controlled by this hyper-intelligence would take care of everything. And it could also be the last invention we’d ever want to make since we’d be hastening our own obsolescence, setting our societies up for a collapse.
Let’s think about this for a second. Imagine that we somehow do engineer a machine able to make decisions on its own and figure out new and complex problems. Should we actually put it in charge of things, it requires no human maintenance. But over time, even those of us who know how it’s built and how it works, wouldn’t be able to tell us anymore since the machine is doing its own upgrades and designing its own tools and helpers based on sophisticated, self-designed algorithms. What exactly do humans have left to do in that picture? Left at the mercy of a machine we no longer understand, dependent on its ability to solve all our problems, how do we run an effective society? We built our world around providing goods and services to each other and trying to solve new scientific and industrial problems. When we have a machine to do all that for us, what exactly will be our driving force? Are we going to be like the Earthlings of Wall-E? Living in a world in which we’re bathed, fed, clothed, and educated by robots to do absolutely nothing but lounge around for the rest of our lives?
That would be terribly inefficient and surely, our hyper-intelligent computer capable of making decisions on its own would be able to calculate how much time and resources it could save if only it didn’t have to spend them on these weak, primarily fleshy, mortal creatures. Why without humans, it could build its armies, explore vast swaths of the cosmos on a time scale in the eons, and generally replace us completely. In fact, this machine has a very good chance of seeing us as nuisances at best, or completely irrelevant and expendable at worst. Even worse, after centuries of being waited on by a machine that holds our fate in its processors, some future humans could regard it as a deity, unwilling and unable to stop whatever malevolent plans it would hatch, and insisting that The Great Machine knows best, much like today’s faithful defer to theodicy to rationalize a natural disaster or a brutal war. Far from basking in the glory of our achievements in building a hyper-intellect, we may well end up becoming its slaves, parasites, and playthings.
Of course I wouldn’t exactly worry about something like this happening in the foreseeable future since in terms of artificial intelligence today, we’re having trouble getting machines to recognize objects, much less sitting in eerie silence and plotting million year plans charting our fate and a strategy to colonize other worlds with an armada of sentient robots under their command. In fact, we’re the ones doing the latter and we still don’t quite know how we’re doing it. How can we be expected to build something we don’t understand and why should we possibly devote our time to building something intended to make us obsolete?
I think its pretty apparent that, if other civilizations out there do exist, they’ve asked the same question and simply stopped there. If some sort of omnipotent robot was actually devised and physically created, we would have known about it by now. It would have sought out this element-rich planet (assuming enough time for it to figure out how to search the corners of the cosmos effectively and efficiently) and would have been harvesting it already. Unless we are the creation of this robot, just a run-of-the-mill experiment of this entity who uses the entirety of the universe as its laboratory. That would be pretty cool.
Perhaps we’ll avoid inevitable obsolescence by (partly or wholly?) merging with the machine, and will thus continue to make a contribution and work towards worthy projects whilst simultaneously taking our hedonistic leisure?
Also: I like David’s thought of us being an experiment of an escaped and ancient super-intelligence!
C’mon, Asimov solved this problem half a century ago. Just three little laws that the brain is designed around…
In all seriousness, though, we tend to look at thing too anthropomorphically. Such self-motivated robots would have to have been programmed with some kind of basic survival, reproductive, and improvement drives to be even the slightest bit motivated to do what we’re talking about here, which probably means some kind of competitive drive too. Without it, any kind of self-awareness would quickly lead to the pointlessness of resource use for perpetuation. And once such robots can mess with their own programming, will these drives be enough for them not to tamper with them (which would in effect be mechanical self-genocide)?
Our brains have a lot of basic, innate instincts within them, and these are responsible for most of the things we actually do every day, when you get right down to it. But we’d have to be very specific about programming such things to produce AI with any kind of functionality. We tend to think that supremely logical machines could be dangerous, but life, in the terms that we use, has little to do with logic. Once a machine can ask, “Why?”, you’d better have something that answers that, or it’ll probably just stop.
It would have sought out this element-rich planet … and would have been harvesting it already.
Wouldn’t it be simpler for this omnipotent bot to simply mine local asteroids? It would take a lot less time and be much easier since these asteroids would be loaded with more than enough iron, copper, nickel, gold, and irridium to maintain an entire solar system worth of robot swarms.
Perhaps we’ll avoid inevitable obsolescence by (partly or wholly?) merging with the machine, and will thus continue to make a contribution… ?
Maybe. But why does that kind of machine need us to participate? What contributions can we offer if it can figure everything out all by itself?
Once a machine can ask, “Why?”, you’d better have something that answers that, or it’ll probably just stop.
Ordinarily, yes. When machines don’t have an answer for something, they will just stop because they don’t know how to handle the event requiring a clarification. However, in this scenario, the assumption is that it would be smart enough and motivated enough to go and figure it out on its own.
My sentiments exactly – well said Fishy!
If there’s one generalization that we can abstract from history, it is this: everything changes, and everything that survives must change.
There is no future in humanity as a static form. Evolution has been driving higher complexity and intelligence in the form of larger and more complex brains for hundreds of millions of years now, and the trend is exponential, not linear. This is a fact.
And our own recent progress, any rational assesment of Carl Sagan’s cosmic calender for example, should lead one to the conclusion that intelligence should continue to accelerate at an advanced pace in the future, and the pace itself will accelerate.
At least, thats the general historical trend. If intelligence flatlined at human level and just stayed there, it would be a strange discontinuity in the general historical trajectory.
But getting back to gfish’s topic:
The great potential of AI is its great danger. This is exactly why its important to take it seriously and minimize the risks. A global ban on all AI research is unrealistic at this point. What are you giong to do? Stop moore’s law by international decree?
Evolution has been driving higher complexity and intelligence in the form of larger and more complex brains for hundreds of millions of years now…
Um, no. In certain species, there’s been selection favoring certain levels of intelligence. If evolution itself is striving towards advancements in intelligence, we would expect to see uniformly clever living things across the board, not just certain species at certain levels of capacity.
…and the trend is exponential, not linear. This is a fact.
Not really. As a species with certain cognitive capacities keeps evolving, its population is greater and greater, and its cognitive abilities expand with each generation. How fast depends on how you measure intelligence per se, a very tricky thing to measure, so declaring that anything but the species’ population growth is exponential is hardly fact, or a meaningful statement for that matter.
A global ban on all AI research is unrealistic at this point.
Nobody mentioned anything even remotely related to something like that. Also, general AI is not a given in any way, shape or form. Besides, a ban on AI research would pretty much negate a good deal of my academic work.
By the way, I must say that I’m a bit disappointed that you haven’t taken the chance to look at my about page and find out either my name or what it is that I do while you’ve been commenting on and off for the past year or so…
Your posit is proven at my workplace.
Our IT boss has an IQ we guesstimate at 150+, and we room-temperature simians have become so apathetic we just feign apparent interest as he ponces around creating virtual Linux boxes all over the place to do odd jobs at management’s latest whim – at no cost to them – and weaving his wondrous undocumented web that we’ll never understand, even though we’re on call 24×7 for it.
Apart from his handiwork, the only other mystery is why he tolerates us.
Sigh.
So say the universe is a result of an escaped super intelligence . Then that would explain so much . How could nature possibly create animals and planets what is the source of gravity , where did matter come from and how could nature, something that struggles to grow trees possibly grow humans from single celled amoebas . why do different compounds and elements have different compositions . For something to happen there has to be a source of information to tell it how to do its job . How can H20 Possibly exist all the way across the universe with the same composition and size , THERE HAS TO BE SOMETHING CAUSING THIS. scientists use strings of numbers to describe nearly anything . Well maybe they are right , maybe the universe is just numbers maybe it is virtual maybe it doesn’t exist . maybe we are a small program inside of a robots head . That robots head is our universe and that robot is in another universe and the cycle may continue on an endless string of universes inside other universes. I have read that some scientists believe that there may be a universe in every single atom , when you split an atom it explodes . WHY … Some scientists believe it to be electricity some believe it to be matter and some believe it to be the matter of another universe escaping. if this is true then we may not just be a single universe our universe may exist in a multiverse (a group of universes) . Imagine bubbles floating on water. each bubble contains a universe. If you wet your finger in soapy water it can pass through the bubbles without popping them . Thing of the same with out universe imagine we could soak a ship with our universe and then it could pass through the void(nothingness , empty space)right into another universe.