Edge 2015: What about machines that think?

I can add that the Intelligence Design Labs and its theory of operation (in the Theory of Intelligent Design my name links to) is for modeling the control freak behavior that easily resorts to deception to control something that it wants to control. Making it honest is extra work, more like trying to make bed bugs stop purposely hiding under mattress tags and their disgusting sneaking around at night. Being good at being deceptive has long been more like the rule for controlling a food source that is also intelligent and can control back by trapping or exterminating them. We are also that way. Humans bred intelligent cows and other farm animals that are smart enough to help tend the fields but better be careful where prodding or they control where you land after swiftly teaching not to do that again. Social amoebas (slime molds) control their most tasty host cells by including the best with them when it is time to weather out a dry spell inside bulbs called fruiting bodies they together built around them. It's hard to say which controls the other because without us or amoeba many kinds of living things would not even exist. A cow might have to take frequent trips for milking, but with humans and now robotic milkers to take care of that for them and acres of food all to themselves it's hard for a well tended cow to ask for more, from us their servants. As long as they don't learn why they have giant udders then demand udder reduction leading to protest we're all set in regards to this deception they don't know about, sneaky us. What I have long studied is what makes the way we behave no surprise we dance, sing and spend all kinds of money on sports games to see spectacular muscle movements of one kind or another. It's then possible to apply that intelligent behavior to a trinity of behavior levels, where 2 are intelligent while the other does not need to be for the three to work together to cause human level intelligence and its associated consciousness. As a result a part of us is estimated to be billions of years old, today. Our DNA holds those memories, though brain stored memories from each lifetime we had along the way did not need to be retained for us to still be here right now. The theory is then so far into what the ID movement has been talking about it sounds like religion being slipped into schools coming in "under the radar" even though it's no fault of mine that what the standard naming convention for science calls "intelligent cause" is thereby being explained. The power of science itself made it possible to control the destiny of the Theory of Intelligent Design and not even the world's top scientists can stop me/us! So of course it's hard for someone like myself to pass up a chance like that, even though the responsibility to not do harm with that power is tremendous and getting there takes patience. Gary S. Gaulin

Gaulin:

Human intelligence does require motor systems for speech and such, and other cells in our body (besides the brain) behave like neurons so I must say that yes Mapou I could have been more precise.

That was not my criticism.

It’s still only simple logic that requires a model of human intelligence to behave the same as a real human or else it’s not really a complete model of human intelligence.

It may be simple logic to you but not to me. Again, intelligence is not motivation but is at the service of motivation. If you have a copy of an adult human brain and the ability to have it pay attention to the important things it has learned over the years, then the copy will probably behave very human-like even though it will no longer have the ability to either like or dislike new things. It will ignore them. However, if you start with a synthetic copy of a baby's brain, it will never learn to show appreciation for things like music and the arts. The reason is that it is impossible to have preset likes and dislikes for patterns in memory. Why? Because these patterns cannot be known beforehand. Motivation for a thing cannot precede the knowledge of the thing. This is the reason that the brain alone is not enough to account for human behavior. Animals are different because they only have a few preset motivations for behaviors having to do with simple goals such as avoiding pain and seeking pleasure. Being sneaky or deceptive is a normal consequence of goal-directed behavior. In humans and animals, it is always done as part of achieving a goal. In animals, the goal always has to do with maximizing pleasure and avoiding pain or discomfort. Again, these can be done with simple pre-wired pain and pleasure sensors. This is the basis of what is called reinforcement learning. However, it is not going to get animals to appreciate art, music or chess playing. By contrast, children are attracted to music, beauty and the arts without being conditioned for those things. It is all very simple really but rest assured that my goal is not to change your mind. I don't really care one way or another. I'm just thinking out loud. Mapou

Human intelligence does require motor systems for speech and such, and other cells in our body (besides the brain) behave like neurons so I must say that yes Mapou I could have been more precise. It's still only simple logic that requires a model of human intelligence to behave the same as a real human or else it's not really a complete model of human intelligence. The only reason machine intelligence is not as deceptive as humans is that full human level intelligence has never been modeled. Mice and other animals are also sneaky and deceptive, not just humans. Gary S. Gaulin

Gaulin @16:

If the machine is in fact an accurate model of the human brain then I would expect it to lie, cheat, be sneaky, and demonstrate all other deceptive traits that humans express.

In my opinion, intelligent machines will lie if they are conditioned to lie via known behavioral techniques such as classical and operant conditioning. Unlike humans, once a machine is properly conditioned, it will not depart from it. Intelligence is subservient to motivation, not the other way around. The reason that humans often rebel against their upbringing is that they have free will, i.e., they can create their own motivations (likes and dislikes). Machines cannot do that. Sure, a machine can be programmed to create random likes and dislikes but this is not free will. Giving a machine the ability to create random motivations that go contrary to its existing conditioning is a solid recipe for failure. PS. Obviously, there is more to the human mind than just the brain. Mapou

Godel, as reserved and mild mannered as he was, was a bit blunt in his assessment of the situation:

“Either mathematics is too big for the human mind or the human mind is more than a machine” ~ Kurt Godel “The brain is a computing machine connected with a spirit.” ~ Kurt Godel “Consciousness is connected with one unity. A machine is composed of parts.” ~ Kurt Godel http://kevincarmody.com/math/goedel.html

a few more relevant quotes:

Algorithmic Information Theory, Free Will and the Turing Test – Douglas G. Robertson – 1999 Excerpt: Chaitin’s Algorithmic Information Theory shows that information is conserved under formal mathematical operations and, equivalently, under computer operations. This conservation law puts a new perspective on many familiar problems related to artificial intelligence. For example, the famous “Turing test” for artificial intelligence could be defeated by simply asking for a new axiom in mathematics. Human mathematicians are able to create axioms, but a computer program cannot do this without violating information conservation. Creating new axioms and free will are shown to be different aspects of the same phenomenon: the creation of new information. “… no operation performed by a computer can create new information.” http://cires.colorado.edu/~doug/philosophy/info8.pdf The mathematical world - James Franklin - 7 April 2014 Excerpt: “the intellect (is) immaterial and immortal. If today’s naturalists do not wish to agree with that, there is a challenge for them. ‘Don’t tell me, show me’: build an artificial intelligence system that imitates genuine mathematical insight. There seem to be no promising plans on the drawing board.”,,, James Franklin is professor of mathematics at the University of New South Wales in Sydney. http://aeon.co/magazine/world-views/what-is-left-for-mathematics-to-be-about/ An Interview with David Berlinski - Jonathan Witt Berlinski: There is no argument against religion that is not also an argument against mathematics. Mathematicians are capable of grasping a world of objects that lies beyond space and time …. Interviewer:… Come again(?) … Berlinski: No need to come again: I got to where I was going the first time. The number four, after all, did not come into existence at a particular time, and it is not going to go out of existence at another time. It is neither here nor there. Nonetheless we are in some sense able to grasp the number by a faculty of our minds. Mathematical intuition is utterly mysterious. So for that matter is the fact that mathematical objects such as a Lie Group or a differentiable manifold have the power to interact with elementary particles or accelerating forces. But these are precisely the claims that theologians have always made as well – that human beings are capable by an exercise of their devotional abilities to come to some understanding of the deity; and the deity, although beyond space and time, is capable of interacting with material objects. http://tofspot.blogspot.com/2013/10/found-upon-web-and-reprinted-here.html

Verse and Music:

Acts 17:28 For in him we live and move and have our being.' As some of your own poets have said, 'We are his offspring.' Shoulders - For King and Country http://myktis.com/songs/shoulders/

bornagain77

Although there are many attributes of mind that are not reducible to brain,,,

The Mind and Materialist Superstition – Six “conditions of mind” that are irreconcilable with materialism: Michael Egnor, professor of neurosurgery at SUNY, Stony Brook Excerpt: Intentionality,,, Qualia,,, Persistence of Self-Identity,,, Restricted Access,,, Incorrigibility,,, Free Will,,, http://www.evolutionnews.org/2008/11/the_mind_and_materialist_super.html Six reasons why you should believe in non-physical minds – podcast and summary (Law of Identity: 6 properties of mind that are not identical to properties of the brain, thus the mind is not the brain) http://winteryknight.wordpress.com/2014/01/30/six-reasons-why-you-should-believe-in-non-physical-minds/ Immaterial Mind - video (Law Of Identity) - video https://www.youtube.com/watch?v=720zEnzgTyM

Although those attributes of mind are clearly not reducible to properties of brain, one of the beautiful things about the Intelligent Design/Darwinian Evolution controversy is that it gives us a even clearer indication as to what separates minds from material processes. Namely, anyone who has been around this debate for any amount of time knows that the main thing that most clearly separates what material processes are capable of from what minds are capable of is the creation of functional information. The sheer poverty of material processes to create novel functional information even extends, not so surprisingly, to computers. In fact, Dembski, Marks, and company, have extended Conservation of Information theorems to show that the problem of functional information generation for material processes holds for computers as well:

On Algorithmic Specified Complexity by Robert J. Marks II - video paraphrase (All Evolutionary Algorithms have failed to generate truly novel information including ‘unexpected, and interesting, emergent behaviors’) - Robert Marks https://www.youtube.com/watch?v=No3LZmPcwyg The Law of Physicodynamic Incompleteness - David L. Abel - 2011 Excerpt: "If decision-node programming selections are made randomly or by law rather than with purposeful intent, no non-trivial (sophisticated) function will spontaneously arise." If only one exception to this null hypothesis were published, the hypothesis would be falsified. Falsification would require an experiment devoid of behind-the-scenes steering. Any artificial selection hidden in the experimental design would disqualify the experimental falsification. After ten years of continual republication of the null hypothesis with appeals for falsification, no falsification has been provided. The time has come to extend this null hypothesis into a formal scientific prediction: "No non trivial algorithmic/computational utility will ever arise from chance and/or necessity alone." https://www.academia.edu/9957206/The_Law_of_Physicodynamic_Incompleteness_Scirus_Topic_Page_

The fact that computers will never be able to create 'non trivial algorithmic/computational utility', (i.e. novel functional information), should not be so surprising. This debate was had, and was settled, years ago between Alan Turing and Kurt Godel

Alan Turing & Kurt Godel – Incompleteness Theorem and Human Intuition – video (with Gregory Chaitin) https://vimeo.com/92387854 Quote from video: Turing recast incompleteness in terms of computers and showed that since they are logic machines, there would always be some problems they would never solve. A machine fed one of these problems would never stop (halting problem). And worse, Turing proved there was no way of telling beforehand which these problems were.” The Limits Of Reason – Gregory Chaitin – 2006 Excerpt: “an infinite number of true mathematical theorems exist that cannot be proved from any finite system of axioms.”,,, http://www.umcs.maine.edu/~chaitin/sciamer3.pdf

bornagain77

If the machine is in fact an accurate model of the human brain then I would expect it to lie, cheat, be sneaky, and demonstrate all other deceptive traits that humans express. Gary S. Gaulin

If a computer calculates that machines cannot think, should it be trusted? It is, after all, merely a machine. Perhaps it's intent is to deceive us. Mung

Artificial Intelligence only has to be as real as Artificial Flowers. That's why it's called "Artificial". There is no requirement that an AI model works the same as a Real Intelligence. Real intelligence is primarily studied by other branches of Cognitive Science such as Neuroscience and the Theory of Intelligent Design I represent that is pioneering a new area of Cognitive Science (not AI) where only real (primarily non-neural biological) intelligence is studied. Gary S. Gaulin

I was responding to the claim that things cannot outdo their maker. That's a silly claim. I don't know how we would ever know in a machine is conscious, but I'm pretty sure machines will eventually outdo us in most tasks that require "thinking." This is a recurring theme. First you had programs that could beat most people who know the rules of chess. Then, incrementally, programs beat stronger and stronger players, until they can bet all human players. Pretty much the same thing is happening in poker, which requires behavior that until recently was dominated by humans -- bluffing, risk taking, and so forth. To the extent that "thinking" has useful outputs, machines will master them. The arts will probably be the last to fall, but I suspect that commercial art will take a hit. And I suspect that programs will soon tie up all the useful melodies in copyright knots. Whether AI will always fall into the category of "zombie" intelligence, I don't know. If I had to bet, I would bet that AI will at some point, initiate a legal claim to human rights. Fiction writers have covered this a number of times. Petrushka

Petrushka, I am not sure what point you are making. A cockroach or an ant can do things that we can't do. So what? We have always invented and built tools to allow us to accomplish tasks that we could not do otherwise. Is an infrared or microwave sensor superior to us because the human body does not have infrared or microwave sensors? Why the focus on superiority/inferiority? Mapou

Computers have already taken over complex management tasks, such as the power grid. And stock arbitrage. These are not simple algorithmic computations. They take place in or on computers, but the algorithms are learning algorithms. Increasingly, what they learn is outside what we can know. And this is new technology. Petrushka

I am reluctant to say this, but humans really haven't been that wise. I suspect within 50 years that AI will manage infrastructure and large scale environmental issues. This will not happen because some human genius like Forbin will design a management machine. It will happen because machines will, piecemeal, prove better at managing things. the mastery of poker is a harbinger of things to come. After chess fell to computers, we said, at least they can't play poker. We didn't design them to beat poker. We designed them to be capable of learning. Eventually, machines that learn will be wiser than humans, at least in the narrow sense of being able to manage complex systems. Petrushka

Barry Arrington: Your comments are probably the result of unthinking, reflexive, knee-jerk opposition to any position taken on these pages. We posted a well-known folk illustration of the transition from human-powered to machine-powered work. The song has a tinge of nostalgia to it, but the transition to machine was inexorable. John Henry was the last gasp of heroic human effort in that regard. Barry Arrington: That, of course, gets us precisely nowhere with respect to the seemingly intractable problems facing humanity, all of which call for judgment, wisdom and prudence – the very things we cannot expect as the result of mere algorithmic computation. Agreed, at least in the near term. Long term, it should be remembered that Chess was once considered the ultimate test of intelligence. While it's tough to make predictions, especially about the future, it's reasonable to believe that humans and machines will continue their integration and symbiosis. http://i.space.com/images/i/000/008/931/i02/apollo17-lunar-rover.jpg?1301637968 Zachriel

Zachriel, I am not sure you are serious. Your comments are probably the result of unthinking, reflexive, knee-jerk opposition to any position taken on these pages. But in case you are serious, yes, by all means, we should assign to computers all tasks that require massive raw computing power when running computer algorithms. That, of course, gets us precisely nowhere with respect to the seemingly intractable problems facing humanity, all of which call for judgment, wisdom and prudence – the very things we cannot expect as the result of mere algorithmic computation. Now, you go back to reflexive foaming at the mouth. Barry Arrington

I am an AI skeptic. Conscious AI is always 20 years in the future. However, I think computers will increasingly take over mental drudgery, just as Caterpillars took over earth moving. they will nibble at the edges of white collar work, until not much is left except wishing for stuff. I think this degree of AI could happen within fifty or a hundred years. If AI consciousness happens, it will not be designed by humans. It will evolve in response to the need to make stuff that motivates humans to keep the machines plugged in. If humans continue to have that option. Petrushka

Thinking means different things to different people. In addition, the meaning of thinking may change with the context. It can mean intelligent conscious processing or intelligent unconscious processing. Both types occur in the brain. The activities in the cerebellum, for example, are fully unconscious. Most of the cortex is asleep (both inactive and unconscious) most of the time. We are only conscious of a very small section of the cortical memory hierarchy, the part that we are paying attention to. As a Christian and AI researcher, I am convinced that we will indeed build machines that are much more intelligent than we are. However, they will not be conscious. It is a conceptual disaster to conflate intelligence with consciousness. The mainstream AI community's belief that, somehow, by some unfathomable magic, a machine will reach a point in complexity when it suddenly becomes conscious, is pure pseudoscience in the not-even-wrong category. Their fear of a robot apocalypse during which intelligent machines will become so smart that they will revolt against humans is nonsense. Intelligent machines cannot generate their own goals. We will give them their goals, their likes and dislikes. If intelligent machines become a problem, we will only have ourselves to blame. The hallmark of human consciousness is an appreciation of beauty, music and the arts. Neither machines nor animals have that appreciation. Future intelligent machines will learn over time what sort of patterns humans are attracted to but that's about it. Mapou

Petrushka: Except, perhaps, in physical strength

John Henry, he drove fifteen feet The steam drill only made nine But he hammered so hard that he broke his po' heart And he laid down his hammer and died,... They took John Henry to the graveyard And they buried him in the sand And every locomotive comes a roaring by says "There lies a steel driving man"

Then they came out with the new version steam drill, two-dot-oh. Zachriel

Second, we certainly should know that what is made never exceeds it maker.

Except, perhaps, in physical strength, or perhaps in the ability to calculate, or to play chess, or to play poker, or to play the arbitrage game in the stock market. Except for these and other attributes and activities, what is made never exceeds its maker. Petrushka

Well its 2015 we should have a hover board out anytime now Andre

And if any man think that he knoweth any thing, he knoweth nothing yet as he ought to know. 1 Corinthians 8:2 The arrogance and folly of man. zfan

Let's see what part of no don't I understand? First of all, this whole thing depends on our definition of thinking and intelligence. Is it about making a decision based on a statistical calculation of a set of data or is it about what color to paint a room or what dress or suit a person would like to wear? How does one calculate a favorite color or taste or sound? Second, we certainly should know that what is made never exceeds it maker. If a computer arrives at an answer it is because someone programmed it that way which leads us into another problem. How does one program morality and on what are the rules based? Is it done by an atheist or a Christian or a Hindu? Whoever does the programming his or her own personal philosophy is going to be reflected in the end product so is the machine really an individually thinking entity? fossil