Humanoid Robots?

[Hussar]

From The Japan Times:

30-year robot project pitched
Researchers see tech windfalls in costly humanoid quest

Japanese researchers in robot technology are advocating a grand project, under which the government would spend 50 billion yen a year over three decades to develop a humanoid robot with the mental, physical and emotional capacity of a 5-year-old human.

The researchers believe the Atom Project, inspired by the popular robot animation series "Tetsuwan Atom" by the late cartoonist Osamu Tezuka, would help promote scientific and technological advances in Japan, just like the U.S. Apollo Project, which not only succeeded in landing men on the moon but contributed to a broad range of technological breakthroughs.

Researchers proposing the Atom Project include Mitsuo Kawato, chief of the Computational Neuroscience Laboratories at the Kyoto-based Advanced Telecommunications Research Institute International.

The researchers are aware that such generous spending on robot technology is unlikely in the near future, given the government's severe fiscal situation. However, they argue that a project of this magnitude would inject much-needed vigor into a nation depressed from years of economic slump.

Unlike cartoonist Tezuka's "Atom" character, known as "Astro Boy" overseas, based on an image of a 9-year-old boy, the Atom Project aims to create a humanoid robot with the physical, intellectual and emotional capacity of a 5-year-old that would be able to think and move on its own, the researchers say.

As exemplified by the Asimo robot developed by Honda Motor Co., Japan's technology in this field has reached a level where its robots can be programmed to approximate the way humans walk. However, their physical ability is still no match for a 5-year-old human.

An expert in brain science, Kawato developed the DB, a humanoid robot that can imitate human action, based on his knowledge that the cerebellum plays a key role in learning and memory. However, the robot has managed to learn only 24 kinds of human action over three years.

"Most of today's robots operate with a program written by humans. In order to develop a robot that can think and move like a 5-year-old, we have to first understand the mechanism of how human brains work," Kawato said, admitting the difficulty of his project. "That will be equal to understanding human beings."

But the researchers believe such daunting challenges, once overcome in the development process, would bring huge benefits in terms of technology and knowledge.

The U.S. government injected more than $20 billion in the Apollo Project -- the equivalent to 7.2 trillion yen if calculated under the 360 yen per dollar exchange rate at that time.

People remember Apollo mostly for the 1969 landing on the moon, but the project resulted in technological advancements in a broad range of areas, from computer technology to development of cordless and energy-saving electric appliances, to the fluorination technology used in frying pans.

"We do not know what will come out of (the Atom Project), but the important point is that we have a target for which we can devote our energy," Kawato said. "We need a project that may not bring immediate economic effects but will lead to future technological development."

Takayuki Furuta, chief of the Future Robotics Technology Center at the Chiba Institute of Technology, who took part in the development of morph 3, a 38-cm-tall humanoid robot, also said he hopes the Atom Project helps create a future industry involving both the academic and business sectors.

SO, just take for a second that they are successful. They create a robot with the physical and mental capacity of a 5 year old child (which is pretty damn impressive). What impacts will this have on our society? Is this a goal we should be pursuing?

[Merkuri]

I find this fascinating from a programmer's standpoint. Creating a robot that could emulate a 5-year old's emotions and thought-processes would be an awesome step in understanding the human brain. And if we were to understand the human brain better we could doubtlessly improve our current computers. I mean, the capacity of the human memory is almost limitless. According to theory, you never forget anything, you just lose it among all the other memories. But it's still there and can be brought out by any number of triggers such as a unique scent relating to the memory or hypnosis. To emulate that memory structure and combine it with a more reliable linking system like we have in today's computers would be a huge jump in computing.

But then there's the moral issues. When does a computer become sentient? Does it ever become sentient? If we create a robot that behaves exactly like a 5-year old child it will be very tempting to anthropomorphize it and believe that it is a conscious living child. Scientits will say that it is not sentient because processors and circuits, no matter how complicated, are not self-aware and can never be conscious, though they may be able to emulate it. But we understand so little about our own consciousness, so who's to say that circuits can never be conscious? Maybe consciousness doesn't have anything to do with the brain or with cells or even with life at all. We really have no idea. Based on near-death experiences where a person was clinically brain-dead, yet they have memories of what went on during this brain-dead state (usually seen from an out-of-body perspective), some scientists now believe that "consciousness" resides in the cells of a body rather than the brain. If it can be found in cells, why can't it be found in circuits?

I think that developing a robot like this is inevitable. There will always be somebody who wants to do it, and whether it is a government funded institution or a university or some genius in his garage it will happen some day, so I say go for it Japan. Who knows, we may never be able to do it. But we might as well try.

[CosmicHorror]

Well, I support the idea, but I'm scared that my computer might talk back sometime...

[Sharess]

Hey i'm all for ti too but what if your pc is having a bad day and goes all pysco I can see it now a real matrix

[Merkuri]

I suspect that once they get it working they'll only incorporate certain aspects into personal computers, etc. I can't imagine why somebody would want their computer or their car to have emotions unless they were really lonely. That can only cause problems, and if they do implement anything like that I'm sure there'll be a shut-off switch for it.

[Shura]

The idea of sentient robots is quite disturbing. And probably best if never realised, as we have enough trouble dealing with organic life anyway.

Thats the practical view, but part of me is thrilled at the idea, just to see how far it ca be pushed...

The concept of golems comes to mind. As Vimes said of the golem situation in Feet of Clay-'it was like a bunch of shovels meeting for a chat'

[Hussar]

Heh, great read.

On another point, right now we have a great deal of difficulty with issues like abortion. Is a fetus a "life" or not. Imagine you create a sentient machine. It passes all the Turing tests that you give it. How can you turn it off? Is that not murder? What rights would a machine have? If it has none, then why do we have any?

Or take it a step further. Another science fiction idea that has definite possibilities is uploading personalities. Copying your mind onto a computer. What rights does that copy have? Is it human? If it is a perfect copy of someone's mind, can it be baptised? Married? Enter into legal contracts?

[EK - Shadow of Death]

Sorry about the juvenility of this post, but this sounds exactly like the Matrix

[Merkuri]

Sorry about the juvenility (is that a word?) of this post, but the idea of AI with emotions, etc came long before the Matrix. Although it was an aweosme movie, Matrix was by no means original.

Not to hyjack the thread... I stop now.

[Merkuri]

Hussar, I will bet money that no matter how advanced this robot becomes or how many emotions it can display or Turing tests it can pass, the scientific world will NEVER label it as sentient or living. Because they all know, especially the scientists that worked on it, that inside it is all circuits and processors. It is doing exactly what it is programmed to do, and what it is programmed to do is COPY human emotions and thought processes. In order to be conscience, the robot must be able to be aware of itself, but processors, no matter how fast or what program they're running, cannot "percieve" anything, they can only process data and spit out results. It may be able to fool people into thinking it is alive and aware, but that's all it will be doing, fooling people. Scientists will never label a robot conscious until they determine what makes US conscious, and I doubt that will happen in our lifetime, if ever.

Thus, if scientists have any say in the matter, it will never be given any rights. A fetus will one day become sentient and conscious, the big question is just when. There is no question that the ability for consciousness is there, developing. Robots, however, are still pieces of metal and plastic executing pre-determined programs.

To phrase it another way, I suppose you could say that robots have no free will. We feed into them a set of instructions (albeit a very complicated set) and they follow those instructions exactly. If you took two robots with the same code and gave them the exact same stimulus you would get the same exact result. It doesn't matter that the robot cries when you tell it that you don't think it is self-aware. The programmer told it to do that. Anybody with the code and enough time on their hands could predict what a robot would do in any circumstance. Therefore, it has no free will. And because it has no free will, it will never have rights. It's very existence eliminates its freedom, it won't matter whether we don't allow it more freedom by giving it the right to life and the pursuit of happiness.

[Hussar]

In order to be conscience, the robot must be able to be aware of itself, but processors, no matter how fast or what program they're running, cannot "percieve" anything, they can only process data and spit out results.

Sorry, were talking about a computer or a human brain? There is absolutely no proof of our own conciousness. A brain cannot perceive anything. It can only recieve inputs from its receptors and process that information.

We do have a method for determining sentience and that's called the Turing Test. If the computer passes the test, it's concious.

[Merkuri]

Apologies, Hussar, I was thinking of something else when I read "Turing Test." But I still hold by my opinion that we cannot declare something conscience until we understand was makes us sentient. The Turing Test deals with "intelligence". "Inteligence" and "conscienceness" can be defined in very different ways. We may be able to prove that a computer can reason like a human being, but can we prove that it can percieve things like we can? No, because we don't understand it ourselves. And I believe that no matter how smart something might be, what really matters is if it knows that it has no rights or that it is in pain or that it doesn't like that color. People don't fight for animal rights because they're smart. They fight because they believe that animals can percieve things like we do, that they can suffer. What does it matter if a robot has no rights if we know it can't suffer? And no Turing Test can prove whether a robot can suffer.

By the way, I'm assuming (like most other people do) that since I can percieve things that other humans can, as well. Yes, I know we can't prove it. But it's a lot harder to apply that reasoning to something we built and something we know the inner workings of.

[Hussar]

they believe that animals can percieve things like we do, that they can suffer. What does it matter if a robot has no rights if we know it can't suffer?

But, how do we know that an intelligent machine cannot suffer? If it says that it is in pain, if it says that it is bored, how can we prove otherwise? Descartes started this train of thought by saying that although a machine may say that it is in pain, it cannot really know what pain is. But, that begs the question, how do we know what pain is? What is painful to you might be pleasurable to me. To simply assume that all people percieve in the same way is a fallacy.

Of course, this is one of the main goals of developing a sentient machine. Hopefully, through the process of development, we will come closer to answering some of these questions about ourselves. Finally we would have an "other" to compare ourselves to and come up with some better answers.

[Merkuri]

I agree with you that we can't know whether machines can suffer, I'm just saying that it's easier to asume that they can't, especially since we know that it's only doing exactly what it was programmed to do. When the room gets cold, we feel cold and so we might say we are cold. However, a robot doesn't need that intermediate step. It's already possible for us to create a robot that senses the temperature and when it is below so many degrees the robot says that it is cold. It's easy to assume that the robot doesn't have the intermediate step of percieving the cold because it doesn't need it. It has no choice but to say that it is cold, whether it percieves the cold or not. I agree that we may never know for certain whether it percieves the cold, just that it's easier to logically prove that they don't.

IMHO, we will never produce a truly sentient robot for the reasons I've said above. We can't give a robot free will. Even if we program it to make choices, it needs data to make those choices and any program given the same data will make the same choice. Even if we try to program randomness into the robot it's not really possible. I've said this elsewhere on this site, but no random-number generator or dice-roller is ever truly random. It takes one number, usually the current time or something else semi-random, and performs arithmetic on it to get your "random" number. But that's still input and output. The same input always equals the same output. Unless we create a radically different system for our computers we will never be able to escape this limitation. The same input will always equal the same output. No free will. No sentience.