 |
Stork: Are there pathologies associated with that?
Pinker: There are even pathologies in one's sense of self. There are people, for example, who are, after suffering brain damage lose their sense of where their own body ends and the rest of the world begins. They may look at their own leg and think someone else's leg as been put in bed besides them. People who's brains have been severed, the cerebral hemispheres have been disconnected to treat epilepsy, will sometimes do things because one hemisphere has received an instruction, and then they will make up a story as to why they did it because the part of the brain that makes up stories and that can talk is in the left hemisphere (sometimes the part that controls behavior is in the right hemisphere). So, we have a person that does one thing and makes up a story to rationalize his actions after the fact.
Stork: Hal had a pathology, perhaps. He committed murder. Would you like to speculate on what kinds of cognitive systems could have led Hal to commit murder?
Pinker: It's a complete mystery why Hal would have committed murder. And, I think this is a projection of fears about human evil coming along with human intelligence that we have projected onto Hal, because why do people commit murder. Well sometimes there are rivals that would be convenient to dispose of. Humans have a large set of motives, some of them ignoble, such as rising to the top of a status hierarchy, or competing for fame or mates, or getting rid of someone that has insulted you. It's not clear why a machine that is designed to control a space ship would all of a sudden, out of the blue, have these motives, which make sense in terms of creatures who have evolved, but don't make sense in terms of an intelligence that was designed by a designer for certain functions. I think this is a deep seated human fear. It goes back to Pandora's Box, Prometheus stealing fire from the Gods, Faust's bargain, even Pinocchio being brought to life then turning out to be a mischievous. We think of the human condition as one as being both smart and sinful, and it's hard to tease them apart. And, we think of whenever an artificial system becomes smart, the sinfulness must come along as punishment. Where in as fact, our ignoble impulses like lust, or greed, or jealousy, or rivalry, and so on, have been specifically installed for an evolutionary purpose, namely he who leaves the most descendents, she who leaves the most descendents wins the game, and is most likely to have been the ancestor of all those that we see around today.
Stork: Actually in the film, Hal has two competing drives. One is to keep the mission going, the other is perhaps to protect himself. He's scared that the astronauts are going to disconnect him. So that might be a justification without being really ignoble?
Pinker: Hal seems to have these competing motives going of keeping the mission going and protecting itself. But, self-protection, is itself, an adaptation that evolved in animals for a certain purpose. And, unless the designers of Hal anticipated that it would be likely to be disconnected by some hostile force. There is no reason to build self-protection into the computer. It's expensive. It's extra software. It takes up space. It might get in the way of running the mission. There is no reason to think it comes along for free with the intelligence. In fact, even in human, in animals, we don't protect ourselves at all costs. A parent will sacrifice himself or herself to protect a child. Sometimes allies in battle will do the same thing. So self-preservation is a rather special gadget which evolved in the brain. It's not something that comes along for free when you get something smart.
Stork: Are there any questions you would like me to ask, any ear candies, any particular points, specially centered on AI, Hal, 2001, cognitive science, that I just don't know to ask.
Pinker: The main one that struck me is the whole genre of computer runs amuck films seem to be based on this idea that self-preservation, or grandiosity, or malevolence, or evil are just part of the package of intelligence, and from a cognitive point of view, from a biological point of view, that's the part of the film that makes the least sense; that is a projection of our fears about human condition, and least reflects the realities of artificial intelligence.
Stork: You had given a characterization of how human minds evolved and how computers evolved. Can you tell us in a succinct way your theory of how the mind has evolved and why we're good at what we are good at?
Pinker: Human minds evolved as a modification of the designs of minds of other animals, the primate that a common ancestor to chimps and humans, and we inherit, to begin with, a lot of baggage from primates: a visual system, a way of controlling arms and legs, certain social emotions like fear and status and sexual desire. On top of that, I think what's special in human evolution is that we occupy the cognitive niche in nature. That is, we have figured out way of outsmarting other animals and plants. Their defenses against getting eaten. By building mental models of the world, figuring out how the world ticks, then exploiting it to our advantage. Figuring out how to extract a poison from one animal, dipping a spear into it, and using it kill another animal. Or driving game by making a lot of noise in settle places till it can be ambushed or hurtled over a cliff. Or digging a trap and covering it with branches. We use our intelligence to get around the superior reflexes of other animals. The other thing that drove that evolution of the human brain was social cooperation, that a dozen people coordinating their behaviour and acting together can get a lot more done than one person, Robinson Crusoe having to do everything on his own. And, so I think, we evolved language in order to share expertise and coordinate our behavior. We evolved emotions, like friendship and solidarity, plus emotions like anger and guilt to prevent ourselves from getting exploited by the whole coalition. So if you combine the technological know how that all humans have and the social know how of how to cooperate without getting exploited, I think you see you the two pressures that led to the expansion of the human brain and the evolution of human intelligence.
Stork: Why are we the only intelligent system?
Pinker: Why are we the only intelligent system? And why was it some primate in the Meioscene that turned into a human as opposed to a cat or a bird, and no one really knows the answer because it only happened once, and so we don't have repeated experiments to tease apart the factors. But, I think there are a few plausible guesses. One of them is that we are social and our ancestors were social. All of the great apes are. And, in general, social species are smarter than solitary species. And, its creatures like dolphins, canines, cats, parrots, that within their group tend to be the smartest, and they are the ones that hang out in groups. Possibly because you need intelligence to coordinate behaviour, possibly because you need it not be exploited. Also, our ancestors had hands, not originally to make tools, but to climb trees. And, once you have hands any gains in intelligence can be translated into an advantage in the world. You can manipulate the world with hands that you can't if you just got paws. A third thing might be that our ancestors probably hunted and ate meat, in general hunting species tend to be smarter than vegetarians, partly because you might need more brainpower to subdue an antelope than to subdue grass, for example. But the other is that brains are expensive. Our brains occupy 2% of our weight but use 20% of our nutrients, and to stoke this hungry brain you need concentrated sources of nutrients which you find in meat more than in fruit or grass. And, its possible that the advanced visual systems of primates helped give a kind of kick-start to intelligence. About half of a monkey brain is devoted to vision, and vision gives us a very detailed understanding of the physical world, three dimensional arrangements of objects in space, and the colors and visual textures of objects give an understanding of the stuff that it is made of. And, it may be that if you have a cognitive framework in place that registers the structure of the world, what things are made of, you can transfer that to more abstract problems. You can solve math problems by visualizing graphs, for example. That may have been another tool that was in place that helped human intelligence get going.
Early in 2001, you see a chimp-like creature, presumably our ancestor, first figure out that you can pick up a bone, use it like a tool; no sooner does he do that than he realizes he can whack a rival over the head with it. And, in fact, weaponry has been one of the main uses of human tools, at just about every stage. We know that spear points and bow and arrows are quite ancient. We know that whenever a human group invents a technology, such as a poison tip spear that is good at felling an animal, when they get into an argument they won't hesitate to use it on another human being, quite effective. We see that in the central dilemma of science and technology, even today. Science and technology are ways for us to manipulate the world more effectively, to get what we want. If something that someone wants is hurting someone else, then technology is going to allow them to do it more effectively. And, the choices that we face as we develop science and technology are the benefits of being able to control our world more effectively with the costs of it getting into the wrong hands of people who may have malevolent uses for that technology.
Stork: One of the grand sweeps of the film 2001 is the evolution of intelligence, from the discovery of tools to Hal, artificial intelligence, to this enigmatic Star Child that appears at the end. Would you like to speculate on the large-scale evolution of intelligence?
Pinker: One of the themes of the film was that there was a natural progression from the biological evolution of intelligence, from apes to humans, to the evolution of human technology, to space travel, and then to some very strange and mystical later stage in evolution. I think that one theme that doesn't fit so well with our current understanding of the development of the mind and brain; that the brain isn't the culmination or goal of evolution. It's just a gadget that worked for a particular primate in a particular ecosystem. Evolution works simply by generating random variance, and the ones that are most successful in surviving and replicating, survive and pass on their traits. It's not something that pushes towards greater complexity, or intelligence, or spiritual enlightenment in general. And indeed, of all of the millions of species that exist on earth, all but us have done quite well without this highfalutin intelligence. Also, once we evolved our brains, what we did with it didn't necessary continue the process of evolution that led to the human brain, because when we invent thing, we don't just generate millions and millions of combinations and see which ones work, we play things out in our mind's eye, we try out possibilities in our head, we share information with other people, we accumulate the discoveries that other minds have made thanks to language and social interaction, all of them are options that are not available to biological evolution and they make the process quite different. And the idea that this will somehow be extrapolated and we reach some other level will be very hard to make sense of in terms of the history of human technology or the biological evolution of the brain.
MOC: Human consciousness and intelligence seem to have arisen out of this mass of neurons and synapses. Is there any reasons to think that this can or can't be duplicated by artificial means?
Pinker: One of the great mysteries of human versus artificial intelligence is where conscious comes from? Consciousness not in the sense of being to report some things and not others, I mean, I am conscious of what I plan to do and what I plan to say, but I am not conscious of the muscle movements or the image on the back of my eyeball. That we can see analogs to in computer systems. The hard problem of consciousness is why does it exist at all? Why does it feel like something to be me, here and now, seeing blue things, hearing noise in the room, subjectively, in the present tense, in the first person sense? And there, it's a mystery of why 100 billion neurons just sending electrical discharges and chemical packets should result in it actually feeling like something, feeling alive to be that brain. We don't know how that happens. Then if we were to develop an artificial system that duplicated everything that a human did, that produced human sentences, that were sensible, that reacted to events the way that people did, would that system be conscious? Would it actually feel pain when you kicked it or would it simply ouch because it was programmed to do it? And, you get very different answers if you ask different people. Some people will say it's just meaningless to ask the question. Once we look at how you can tell whether someone was conscious, any computer would pass that test, it's just chauvinistic to say that we feel pain when say ouch, but the computer doesn't feel pain when it says ouch. There are others who say its obvious that the computer feels pain and that the computer doesn't, just because consciousness has to arise out of neural tissue, the way bile is secreted by the liver, consciousness is secreted by the brain. I can't think of any way of distinguishing between these very two strongly held theories. My own sense is that the human brain may not be capable of formulating an answer to that question. We may be able to ask some questions that we simply can't grasp the answers to in the same way that other animals can not figure out prime numbers or color blind person may not be able to see the difference between red and green even if intellectually they know there is such a difference. I can say one thing, it may be that our befuddlement at the question "how does information processing in the brain give rise to feeling" is like our befuddlement when we ponder the mysteries of quantum mechanics. "How can the damn particle go through two slits at the same time?" Or relativity and cosmology. "What was in place before the big bang?" The physicist say time didn't exist before the big bang. I can grasp that intellectually, but there is still part of me that is saying, well okay, but what was it like before that, even though it is a meaningless question. And, it may be "what does it feel like something to be inside a brain processing information" may be just as meaningless a question, but one that we can't help think, having human brains.
To learn more about Steven Pinker visit:
Or
|
