Japan Goes Robotic
by
Suzanne Hadley
on Mar 6, 2008 at 8:09 AM
Over Christmas break, my dad and I watched I, Robot. In the film, robots, which have been programmed to be harmless and helpful, begin to violate those rules and turn against humans. Will Smith has to try to save the day. For the most part, the robots in the film are portrayed as creepy and menacing. But in Japan, robots are not only a part of life; they are its future. AP reports:
There are robots serving as receptionists, vacuuming office corridors, spoon-feeding the elderly. They serve tea, greet company guests and chatter away at public technology displays. Now startups are marching out robotic home helpers.
They aren't all humanoid. The Paro is a furry robot seal fitted with sensors beneath its fur and whiskers, designed to comfort the lonely, opening and closing its eyes and moving its flippers.
But robotics in Japan isn't just for fun. The country is depending on robots to solve some of its biggest problems.
For Japan, the robotics revolution is an imperative. With more than a fifth of the population 65 or older, the country is banking on robots to replenish the work force and care for the elderly.
My first reaction to robots that act like humans is suspicion, but the article explains why the Japanese psyche is more open to robots.
Japanese are also more accepting of robots because the native Shinto religion often blurs boundaries between the animate and inanimate, experts say. To the Japanese psyche, the idea of a humanoid robot with feelings doesn't feel as creepy — or as threatening — as it might do in other cultures.
The Japanese government is counting on 1 million industrial robots by 2025. The article closes with a rather creepy (in my opinion) sound byte from Hiroshi Ishiguro, a robotics engineer at Osaka University. Ishiguro favors making robots that look just like human beings.
"One day, they will live among us," Ishiguro said. "Then you'd have to ask me: 'Are you human? Or a robot?'"
1. Amir Larijani had the following to say on Mar 6 at 8:19 AM:
Well...that will create more work for robotics engineers and technicians.
Lots of opportunity....
2. NeedACatchyName had the following to say on Mar 6 at 9:01 AM:
There's a lot of misconceptions about robots these days. Generally speaking, people tend to try and anthropomorphize robots too much. When you say "robot" most people think about a walking, talking metallic humanoid (something like Rosie the maid on the Jetsons), when really no currently existing robots are anything like that. Really, a "robot" is just a term for a series of tiny motors and appendages controlled by a computer. A robot is far more likely to be a computer controlled mechanical arm welding a car frame than a metallic, humanoid looking contraption that can interact with you on a conversational level. Heck, there's a robot in the office a few doors down from me--it's a computer controlled mechanical "arm" that takes CDs out of a CD label printer and stacks them when they're done being printed.
The "robots thinking on their own" idea that some people seem to have is also far-fetched. A robot is controlled by a computer, so it has all of the same limitations of a computer, and can only do what it's programmed to do. In much the same way that your personal computer lacks sentience and is not plotting against you to take over the world, robots aren't some sort of super-intelligent creations with the ability to think on their own. They're just computers with moving parts.
3. Christina (in green) had the following to say on Mar 6 at 9:05 AM:
Yeah...I can see this in japanese culture...
I remember chobits...
4. Christina had the following to say on Mar 6 at 9:34 AM:
Sounds like an episode of the Twilight Zone. (enter creepy music)
5. Mike Theemling had the following to say on Mar 6 at 10:04 AM:
Like NeedACatchyName said, a robot is simply a machine that does a programmed task.
An android however is a robot which is designed to mimic human appearance and motion.
No surprise that Japan is the main promoter of this type of alternative workforce. 1/3 of their population will be over 65 very shortly with a low birth rate. They've always seem to have had a fasication with robots in their culture. They gave us Transformers, Voltron, Robotech (Macross), and a host of other robotic characters.
I don't think robot sentience or even programmed sentience will occur anytime soon. Even if we do run into a Blade Runner type society, we can always make a Voight-Kampff machine to distinguish between the organics and the mecha.
6. Amir Larijani had the following to say on Mar 6 at 10:07 AM:
NeedACatchyName:
Actually, leading-edge robotics technology incorporates artificial intelligence, which does expand the capacities to include making decisions that were not specifically programmed into it.
Humanoid robots have actually come a very long way, and--I'm not happy to say this--the porn industry is already pushing the envelope of the technology.
THAT, my friend, is a whole different discussion all its own.
7. obewan had the following to say on Mar 6 at 11:07 AM:
I don’t see robots as having any substantial form or “artificial intelligence” for a very very long time. There was a special on PBS that showed a small mobile robot arm that roamed around the floor touching different blocks. It was supposed to simulate some form of learning that was related to “evolution”.
The robot was supposed to learn from its environment, and get smarter and smarter as time went along with no outside help from a programmer. It required a large supercomputer in the basement of the lab that was connected by a radio transmitter.
There were tens of thousands of lines of computer code, and the so called intelligent robot was limited only to knowing the difference between 3 or so different kinds of blocks, and how to move around without bumping into them.
It looked like a boondoggle with a lot of wasted research dollars, and a heavy handed secular humanist slant that attempted to substantiate “evolutionary theory” with a stacked deck created by a group of programmers and engineers who made forgone conclusions about how “learning” and “intelligence” actually operate in the brain. The focus of their “research” was simulation of the synapses. Still, a robot cannot understand the meaning of a concept like beauty.
8. Amir Larijani had the following to say on Mar 6 at 11:54 AM:
Obewan:
On the other hand, AI allows a robot to learn in order to make decisions with respect to the environment and its own mechanical capacities. There are robotics competitions in which robots traverse paths--using AI--in very impressive timing.
While you are correct that there are concepts that will elude the capacities of robots, AI does allow them to have a range of use in manufacturing, aeronautics and space systems, search and rescue, and--sadly--pornography.
I think the evolutionary angle presented by the robotics community is more hype than anything else. At best you are able to simulate learning to a certain degree. On the other hand, that is a far cry from macroevolution. So anyone who thinks AI is a proof for DE isn't thinking this through clearly.
If anything, it makes a case for ID, as neither the robots nor their intelligence capacities required at least one Designer.
That being said, nothing in my comments was intended to make a case for or against evolution but rather the case that robotics technology has taken capacities far beyond the limits of traditional structured programming.
9. Chris had the following to say on Mar 6 at 7:12 PM:
Mike writes:
I don't think robot sentience or even programmed sentience will occur anytime soon. Even if we do run into a Blade Runner type society, we can always make a Voight-Kampff machine to distinguish between the organics and the mecha.
Ah, yes, but what about if/when AI progresses to the point that it can pass the Turing test? The V-K is just a type of Turing test, IMHO.
I'm not saying that would happen, but one cannot always rely on a "test" to prove something. Make something idiot proof, and someone will be build a better idiot......
But to get to the real point: Blade Runner is one of the best movies ever made......
10. JN had the following to say on Mar 7 at 6:32 PM:
We have among us already those who act like these feared robots with no conscience. We should worry more about how we are failing to program our own children.
11. Robin Munn had the following to say on Mar 8 at 7:44 AM:
Science-fiction books and movies have given a lot of people an overinflated idea of what A.I. (artificial intelligence) can do.
Fictional AI reaches the point of creating a being with intelligence indistinguishable, or even superior to, a human being's intelligence. Most portrayals of fictional AI include sentience (self-awareness, the concept that "I exist") and the ability to feel emotion.
Real AI, on the other hand, cannot and will not ever reach that point. Or anywhere even remotely close to it. There are things a three-year-old child does routinely, such as recognize faces, understand language, and create stories in his/her imagination. All of these are tasks that are extremely difficult computers even with the best of modern AI techniques, and the last of those (using the imagination) is utterly impossible for a computer. To use the imagination requires a spirit, something only God can make.
Yes, computers have been created that can beat human grandmasters at chess. But chess is a very narrowly-defined problem that lends itself well to a mathematical solution, and that's what computers do best. Language, on the other hand, is insanely tricky to describe in mathematical terms. Many times, you have to understand context to extract the meaning from a sentence. Consider, for example, the following sentence:
Fruit flies like a banana.
This sentence has two possible meanings, and each the first three words will be two different parts of speech in each of those meanings. "Fruit" could be a noun, or it could be an adjective. "Flies" could be a verb, or it could be a noun. "Like" could be a preposition, or it could be a verb.
Which meaning is correct? You cannot know that from the sentence alone -- you have to understand the surrounding sentences. Are they talking about how various things fly ("Time flies like an arrow")? Or are they talking about the nutritional preferences of various types of insects? To figure this out, you need to have read and understood the previous sentences, and have some knowledge of the outside world in which to fit your observations.
I won't say that an AI could never be written that could understand language well enough to pass the Turing test, because the Turing test merely requires a human being "chatting" with an AI over text chat to be unable to tell if he/she is talking with another human being or an AI. It's possible that someone would be unable to distinguish the AI's from a human being's conversation -- but that would entirely depend on the intelligence of the human reading the chat, and their experience with AI. In fact, this has already happened once or twice -- there have been times when people were chatting with a "bot" (automated conversation software) on an Internet Relay Chat channel and never realized they weren't talking to a real human being. (Much to the amusement of the regular members of the channel, who knew that "alice123" was a bot). In such a case, where it's obvious to most observers, I would consider that the human has failed the Turing test, rather than that the AI has passed it. :-)
The Japanese are certainly making great strides in robotics, including building robots with fairly realistic human appearance. (The proper term for those would be "androids"). The "robot receptionists" mentioned in the article would be an example of this. Those would function well, because they don't need "real" intelligence of the kind that fictional AI would produce. All they need is the same kind of voice-recognition you get on automated phone menus these days. "Hello, welcome to ACME Corp. What department do you want to visit?" "Marketing." "I'm sorry, I didn't quite understand that. Did you say marketing?" "Yes." "The marketing department is on the third floor. Have a nice day!"
However, we will never see "fully intelligent" androids like Data (from Star Trek: TNG). It's simply impossible.
12. Don had the following to say on Mar 9 at 3:35 AM:
Just thought I would throw out that, at least as far as I can recall, the robots in I, Robot didn't turn against humans. They just took their programming to an extreme, and ended up trying to help humanity by ending it. Again, I haven't seen it in a while, but I think that's what was happening.
13. Christina had the following to say on Mar 12 at 9:49 AM:
Don, that seems about right.
AI, from my understanding in developing semi-AI programs, is nothing but creating complex decision making without input from an external source.
Smoke Detectors use AI (I'm being simplistic) in its most basic form - the sensor senses something and triggers a decision Smoke - yes or no? Yes. Sound alarm. It requires no human help in making that determination.
The machines that appear as intelligent and truly amazing are programmed with lines and lines upon lines of complex decision making code that is only as good as the programmer who could think up every possible scenario.
Evolving machinery would be where the programmer encounters a situation where his robot doesn't perform well and modifies the code to handle that situation =p
14. Evan Pederson had the following to say on Mar 21 at 11:07 PM:
There are various categories of programming algorithems, such as neural networks, that do not require the programmer to anticipate every possible situation.
I'm in the camp that believes a robot will never be able to achieve true self-awareness in the way the humans are, because I believe there is a spiritual element to the human self-awareness that we cannot create.
Artificial intelligence is a different issue. At the rate AI is improving, and considering some of the recent advances in both neural understanding and "learning" algorithems, I don't think it's much of a reach to say that AI's that will appear to be at least as intelligent as humans (in specific kinds of situations) are very far off.