I don't see any way we could end up having a skynet-style robot enemy on our hands. This scenario is based upon Artificial Intelligence being developed and "choosing" to destroy us. This probably won't and can't happen. What CAN happen, is we could create a fleet of robots that accidentally destroys us because we type in the wrong command.

Maybe a billion trillion self-replicating nano-robots designed to clean up an oil spill get the wrong message and start consumer anything with carbon in it. But I highly doubt we will see a robot ever make a sentient decision to fight humanity.

Thats because nothing non-organic has ever been the slightest bit sentient. Theres a lot of misleading info about artificial intelligence that gives people the impression that machines are "learning to think." No, they are learning to compute equations faster which makes their reaction time and deductive capabilities appear to be thought-based. But they aren't, they are program based. What is more sentient; a pocket calculator, or a supercomputer? They're both equally un-thinking machines.

But if computers can work ever faster, won't that come to mimic human thought? Or some kind of thought? No, because everything a computer does is predetermined by its programming. The smartest computer in the world will not be able to tell you how to tie your shoes if it isn't programmed to. If the slightest bug hits its programming, the rest of it screetches to a stop. Machines have static programming, and can't appraise situations outside of that. This applies to 100% of all "artificial intelligence" programs or machines that have ever been created.

Real intelligence is the ability to absorb and process new information. No computer in history has ever processed something that it wasn't specifically designed to. Because it can't, and because artificial intelligence is far more artificial than it is intelligent.

In thinking about this, there is only one way I see that such "thought" could emerge. Perhaps is computers were given the ability to constantly reprogram themselves, then perhaps, maybe, some new sort of evolutionary process could spontaniously generate real decisionmaking capability. But that doesn't seem very scientifically sound to me.

Well my vba skills aren't really qualification, but you may want to look at this a bit more. Theoretically, there is no brain process that can't be modeled including non-deterministic outcomes. Given current limitations in brain scanning, computer power etc., we are limited to recreating the effects of only certain brain regions- but the technology and insights are advancing.

Oh, even better- a real example.

That was a predictable response. ;-)

Ha, I was about to write the same post. Not only are we not close to achieving the creation of anything sentient, we don't even know what it means. We have no idea how our own brains work. And figuring out which parts become more metabolically active when we think doesn't count: that's like trying to reverse-engineer Windows using a thermometer.

I would disagree in principle, though I do agree that current technology appears to be a very long way from anything approaching "human intelligence" -- though of course there is no general definition of what "intelligence" is, let alone "human intelligence."

There is dificulty in saying that "everything a computer does it predetermined by its programming" because the programming itself can be written to avoid pre-determination -- for example by using randomization. Not all 'software' is "static", either, I worked with early self-modifying computer programs in the military in the 80s. Though very limited in some respects, such software can be very unpredictable -- even "chaotic". Once again we haven't approached anything like the human brain -- but then would it take even a human brain to come up with a powerful attack on humanity? A virus presumably has no "will to live" and is itself arguably un-intelligent -- but advanced computer systems are being used to map and now even design viruses (or mutation of them), and biological experiments are often conducted in almost completely computerized automatic 'lab sets'. A virus might be considered just a particular type of 'nano-bot'. The real future danger may not be Skynet sending human-like drones after us, but LabNet exposing just one vial of test matter to just one extra round of 'treatment' --and creating the "killer virus" we have been afraid of for so long.

Computers *can* learn to tie shoes. Indeed there is very advanced topological software designed to deal with "knot problems" that humans can scarcely approach. Actually, the "intelligence" is probably more advanced than the physical ability of 'tactile' machines to actually tie the knots.

Computers *can* learn to tie shoes. Indeed there is very advanced topological software designed to deal with "knot problems" that humans can scarcely approach. Actually, the "intelligence" is probably more advanced than the physical ability of 'tactile' machines to actually tie the knots.
==

No, that is not learning, that is programming. That advanced topological software was designed by humans.

Nor are advanced algorithms that incorporate randomness any evidence of thinking or sentience, let alone a will. I'll grant that, hypothetically, sentience could emerge out of such randomness, but I haven't heard anything indicating any progress in that fashion.

The example posted of a "robot scientist" here - <link>

Well, that sounds to me like a calculator that is capable of running its own equations. Thats automation, yes, but the "decision-making" of how to handle the procedures and deal with the results are all pre-scripted. If the robot discovered something new which did not fit its programming, it would not be able to proceed as a human would.

So what is sentience if not an infinitely complex series of pre-ordained programs? Well, it can be partially defined as being able to perceive and interact with an objective world outside your own being. I'm typing on my keyboard right now and looking at a screen. To a machine, it would be simply receiving tactile data through the fingers and photons through visual receptors. There is no way to extrapolate that into the idea that there are other objects in existence that I'm interacting with. Thats the building block of sentience.

No computer, no matter how advanced, has a shred of that. Data comes in, hits the pre-set programming, and the response is determined. In terms of dealing with the world around them, computers are on the protazoan level.

I will only be convinced that true AI is possible if I hear that a computer can reprogram itself on a fundamental level, and thus change the number of pre-scripted responses it can issue to incoming data. Maybe, just maybe that could lead to the spontanious development of sentience due to a less restricted exploration of the world.

I will only be convinced that true AI is possible if I hear that a computer can reprogram itself on a fundamental level, and thus change the number of pre-scripted responses it can issue to incoming data. Maybe, just maybe that could lead to the spontanious development of sentience due to a less restricted exploration of the world.

=======

It strikes me that, depending on how you define "a fundamental level," this is actually a higher standard than humans themselves could pass.

Like all organisms, humans are fundamentally "programmed" also- by emergent properties of the sequencing of proteins in DNA. Make no mistake, this is programming in a literal sense- and altering it will alter the nature and learning patterns of the organism produced. Indeed, the wide variance in animal behavior seems to be proof-of-concept that sentience is a sliding scale, not an absolute state and that computers are already quickly moving along this continuum somewhere between bacteria and echinoderms.

Computer technology went from Eniac to Big Blue (which, incidently unnerved and beat Kasparov with its "human like" approach to chess) in a single lifetime- given the slope of that development, I wouldn't count the calculators out just yet.

Organisms are not programmed like machines are. There is a plasticity to the brain which allows certain parts to become more robust, or atrophy, depending on their use. A professional musician has the musical "parts" of his brain working more healthfully and efficiently than a non-musician. Computers don't work like that. If you leave a certain program dormant for 50 years in a machine, it will run just as well as if you had been running it ever day.

But yes, our DNA does mean that our brain components are hardwired with certain functions and limitations. However, the facts that that programming comes together to create a result that is greater than the sum of its parts - sentience and flexible thinking - has not yet been explained by science. Perhaps we will be able to apply these human qualities to machines once we understand how they are caused in our selves.

Big blue may have seemed human-like in how it played chess, but what was that impression extrapolated from? A series of chess moves, which was determined by software. Big blue didn't actually excercise any personal intuition or foresight the way kasparov may have felt it did. That is simply personification. Big blue was simply very good at crunching mathematical models in it's head. It certainly didn't know that it was playing chess, or have any understanding of what chess was, or that it was engaged with another being, or that other beings even existed. It lacked recognition of all those factors to the same degree as any other computer that has ever existed.

It may be possible to create a form of life out of computers, but again, we don't even know how our own programming allows this to happen in us. We aren't a collection of inputs and outputs; our minds somehow create awareness as a continuous by-product to our own internal computations. AND, that consciousness becomes an additional factor in our computational functions which adds an immense ammount of flexibility to the things we are able to think about. This is the mechanism that matters if you're on the lookout for a skynet or a HAL, and I haven't heard any evidence that it is out there.

Jw,

I think we all agree that computers are not even close to a human level awareness, (even if we knew exactly what software we needed, the calculations per second aren't there yet). However, you seem stuck on this idea that computer programs are inflexible. In fact there are many many examples of functional "narrow ai" that have "learned" to find patterns from complex inputs sometimes independently. The genetic algorithms I linked above are just one way that programmers do this. Another is with artificial neural networks. Obviously we are still at the point where humans create the parameters, but the results are not always predictable even with the same initial hardware and inputs.

I don't know where along the animal kingdom we could classify the current generation of AI's, and in a way its meaningless right now. However, as someone else said if the exponential rise of capacity continues, there is no theoretical barrier to software that is smarter than us.

When comparing machine and human intelligence, humans tend to overstate humans capabilities. IQ is normed so that the average IQ is 100. In the last 15 years the scoring was renormed because averages had fallen to the point the average was 90. So, 90 became 100 to maintain the standard. My point is that far more of humanity exists on a cognitive level closer to machine level than to the ideal most of these posts are using as comparison. The human norm is not Einstein but closer to Moe ,Larry and Curly. It is not so far fetched to make machine intellect top the Stooges

*** In the last 15 years the scoring was renormed because averages had fallen to the point the average was 90. So, 90 became 100 to maintain the standard. ***

Wow, you really have NO idea what the fuck you're talking about.

In reality (planet earth, rather than whatever planet you are currently orbiting) IQ's rise about 0.3 points per year. IQ tests are regularly renormalized to make them MORE DIFFICULT - precisely the opposite of your claim, though not nearly as large an effect.

<link>

Well ok, you've made a good point that there is software out there that can create non-static results. This could be a component of making a thinking machine. And yes, the exponential barrier of computing power requires that we leave open the possibility of just about anything happening.

So on the grounds that computers are a relatively young technology I'll concede that its all possible.

What I was initially responding to was the idea (which gets floated a lot in popular technology magazines like Wired or Scientific American) that we have already begun to create primitive thinking machines, when we actually haven't. When it comes to intelligence, a mind either has a degree of self-awareness, or it has none. The functional boundary that separates those two types of computing systems has not yet been found. But still, an earthworm has a greater degree of awareness (however small), than a supercomputer AI (that has none).

Thats basically what I'm pointing out - that there is some as-yet unquantifiable characteristic to intelligent thought that generates self-awareness. It doesn't seem to me like we have even begun to scratch the surface on how to program such a variable into any system. And until we do, I'm not going to be concerned about t-800's. While, possibilities being what they are, I won't discount the idea that we could generate such thinking machines, I also will continue to consider that possibility that we may never be able to do so because, for whatever reason, it is not a quality that can be invented.

"Thats basically what I'm pointing out - that there is some as-yet unquantifiable characteristic to intelligent thought that generates self-awareness."

Everything we know of as aware so far has the benefit of sensory input bounded by or responsive to the physicality of the organism. The field of vision and tactile feedback will respond based on the actions of the organism, which probably helps immensely.

For more advanced awareness, or consciousness a necessary characteristic is probably some level of recursiveness I don't know of any programs that have been set up to analyze their own parameters yet. So far, I don't think anyone is even trying to create awareness.

Re: Siempre

Ignoring the IQ statement, you have a good point. We are all pretty impressed by these organs that allow us to focus for 1-2 hours at a time, hold 7 pieces of information in working memory, learn new information if its repeated a few times, etc.

Just to throw an idea out in the air, what about an AI rising by mistake? Combine our own experiments with AI with the ever growing complexity and chaos of computer systems...