Along Equal Lines

01-02-2007

With the tin-can versions gone, we're wondering what the new humanoid-type robots will be up to

Researchers are falling over themselves to re-enact The Creation, and people are already accepting robots as friends, companions and lovers. In Japan, household robots number in the millions. Most robots today are things like vacuum cleaners, and the rest are doing some cute, some interesting, and some boring things like carrying on micro-conversations ("What is your name?"), climbing stairs, saying "Whoa!" when pushed, playing football, eating sushi, and serving as receptionists. (Spot which one we made up.)

The mindset towards robots has changed in a surprisingly short span of time. A couple of decades ago, they were seen as curious toys, or as miracles, or as dangerous entities that would threaten our domination of the planet-in short, much ado was being made about them. Today, tell someone "Fitzy" won the annual table tennis championship for robots, and he'll tell you who won the rock-climbing championship.

We're amazed at the intelligent things these once-toys can do. What will they be able to do a decade from now? Thirty years? Will it be like in Bicentennial Man, where they're almost part of society, with the same status we accord to ourselves? Will they be classified as a species? Will they have noisy sex? Just how intelligent can they really get? Etc. etc. etc. Of course we don't know any of the answers, but it won't hurt to take a romp through the questions.

Eye To Eye
Pong, the Sympathetic Robot, was developed several years ago at IBM's Almaden Research facility. It is a moving head that "sits on a table and looks at you with two ping-pong ball-like eyes; once it makes eye contact, it smiles or frowns based on its interpretation of your mood.

"I found it fascinating and creepy at the same time. There is something deeply unnerving about a machine looking at you. If the researchers replaced the ping ball eyes with realistic glass eyes, I'm sure it would send shivers down your spine.

"The idea that a machine can approach humans may be deeply repugnant. Perhaps we do not want to develop thinking machines that can look you in the eye."

That's John Dvorak's 2001 account of his experience with a robot that's just barely intelligent. Dvorak is no newcomer to intelligent creations-he's a world-renowned technology author and columnist. But he already sounds ancient, almost quaint. Six years on, and people are talking very differently, as you'll see.

All Too Human
Take a look at http://snipurl.com/digrob2. "EveR-2 Muse," designed by the Korea Institute of Industrial Technology, has been called the world's first entertainer robot. She lip-syncs, does dance moves, and such; and since people don't like seeing hideous steel structures doing these things, EveR-2 Muse has a realistic body-artificial skin, hair, a pair of you-know-whats, and, yes, artificial eyes, of which we think a better job might have been done-but who are we to complain? Still, those eyes are horrifying, and one blogger says that "that glazed, nightmarish glare speaks for itself." We conclude that you probably wouldn't feel uncomfortable with this one looking you in the eye, but what if "she" had "realistic glass eyes"?

Now, take a look at the movies at http://snipurl.com/digrob1. WT-5 (Waseda Talker No. 5) doesn't quite look human, but from the videos, you'll gather, even though it makes Japanese sounds, that it's alarmingly advanced in terms of speech production. The last video is the most helpful, where the bot mimics a person saying something. Where have all the nice tin-can robots gone?

Then there's professor Ishiguro, who emphasises human-like appearance in his robots: head to http://snipurl.com/digrob3. The page says Repliee Q1expo interacted with people "with impersonating a TV interviewer," which we're taking to mean that the bot was presented as life-like. The first video clears things up, and all of them are worth watching. If only those body movements could have been more fluid! And then you realise you're already comparing the bot to a person, which is the ultimate in flattery as far as the professor is concerned. (Chuckle, chuckle: at the bottom of the page, there's mention of Ishiguro's writhing and someone else's greed.)

For a good listing of humanoid robots, visit

http://snipurl.com/digrob4

. They can do rather more than show off by climbing stairs-and that was just a couple of years ago. Which is precisely what makes us wonder what's round the corner-and even more, in fact, about the distant future. If this were a tabloid, we'd already have been screaming, "They're coming! They're coming!"

Another, obvious point in the case for humanoids is that they would
be more readily accepted as companions and helpers. No-one
wants to cuddle up with a twelve-legged, titanium box

In Our Image
Sebastian Thrun, director of the Stanford Artificial Intelligence Lab, was in the news last year for "Stanley," his robotic car. It drove itself through about 130 miles in the Mojave desert to win the DARPA Grand Challenge for self-driven cars. He makes a point about the Roomba-the most well-known robotic vacuum cleaner-that spells out something fundamental: "I think the Roomba is a mix of curiosity and cutting-edge technology and actual utility. It's not the world's best vacuum cleaner, but it's incredibly cool to have a robot vacuum cleaner."

Coolness is certainly a big driver when it comes to consumer robotics, but why this drive towards humanoid robots? It could well be a multiplication thing, as in coolness multiplied by human-like equals Ã¼ber-cool. It's cool to have a robotic vacuum cleaner, but it would be Ã¼ber-cool to have an android vacuum cleaner! Still, there are more reasons we can think of.

For one, they're cheap labour. A receptionist robot costs a lot right now to hire, but when they become commonplace, it really is as simple as that-cheap labour. Then there's the obedience factor-in fact, the word "robot" comes from something in the Czech language along the lines of "doing one's bidding." Who doesn't like to order people about? And who wants his orders to be turned down with "Nah… I don't think so!"?

Robots just don't say "Nah… I don't think so." You tell, they do.

In terms of following orders, robots do it with precision. You will not see a sloppy robot-and when we get to having them as household helpers, you aren't going to see water spilt on the floor after the dishes are done! It's this very precision that has given robots a bad name-a person who in all sincerity says the time is "2:52" is thought of as robot-like in a negative way. The "solution" could be to throw in some sloppiness into the robot's AI, but that's a subject of debate. Our point here is that going beyond orders being obeyed, we want them to be obeyed to perfection (in most cases!).

Another, obvious point for humanoids is that they would be more readily accepted as companions and helpers. No-one wants to cuddle up to a twelve-legged titanium box.

Some humanoids are designed for study purposes-WT-5, which we've mentioned, was developed to study how the human vocal system works, roughly speaking.

You could think of a whole lot of other motivational factors-their entertainment value, for example. (This might soon be gone as they become more commonplace, though.) But the ultimate motivation in the design of robots that look, walk, and talk like us could well be the collective "high" of having created something in our own image. Of having played God-literally.

Looking To Japan
So make no mistake about it: we're not divining, we aren't guessing. It's all in the news. (See box They're Crazy!) The wave of humanoid robots has already begun.

Japan is showing us the way when it comes to acceptance of robots in our daily lives. In that country, they're not viewed as potential monstrosities: rather, they're gentle, benign creatures. One reason cited by some is the Japanese (Shinto) religion, where every creature-including robots-have a "life force." There are accounts of how robotic pets reduce stress and make life nicer for old folks whose children don't visit them often-in one old-age home, two robotic seals function as pets for the residents, who are delighted to stroke them and feel them flapping their flippers. "Old fogies," you might say. Or, "Why not just give them a dog?" We're wondering about the dog too, but the point is that the Japanese really seem to more readily assign a "human" status to "dumb, mechanical creatures."

The point also is, if it can happen in Japan, it can't just be explained away by Shintoism; it will happen everywhere, sooner or later. And not just robotic seals and dogs: as the whole humanoid thing catches on, who's to say that old (or young) people, for instance, won't accept robots as "real"?

Some of the latest robotic creations are anything but made of tin, and sound anything but metallic. Pictured here are Repliee Q1expo (left) and EveR-2 Muse (above)

What They're Saying
Hans Moravec, old-time researcher at Carnegie-Mellon University: "The effect of robots clearly has implications for the economy. Social security will have to be expanded, introduced at lower and lower ages, till essentially everyone lives on social security. The taxes will be paid by fully-automated businesses run by robots. And human beings will have to deal with the problem of excess leisure." (What would you do when there's nothing to do?)

Ron Arkin, roboticist at the Robotics and Artificial Intelligence Group at LAAS/CNRS in France: "I would expect people to form psychological attachments with these artefacts. We already do so more or less: to cars to video games to AIBOs. …Would you want your daughter to marry a robot? What would the church think of such unions?" (This one might sound ridiculous, but think about your psychological attachments to artefacts.)

Daniela Cerqui, social and cultural anthropologist at the Institute of Sociology and Anthropology of the University of Lausanne: "I am afraid that the long term future we are building will have no space left for human beings." (To do anything useful, we presume.)

We're not saying the views above are indicative of Truth. But they highlight what the thinkers are thinking right now: that it's a foregone conclusion that the bots will be amongst us-in growing numbers-now on.

They don't have as much intelligence as bacteria and they never will"

Noel Sharkey, Professor Department of Computer Science, University of Sheffield.

They're Crazy!
According to a recent study by the British government, robots could one day demand the same "citizen's rights" as humans. If granted, countries would be obliged to provide social benefits-including housing. Don't laugh at us-that's what the officiously-called "study" says!
A "monumental shift" could occur if robots develop to the point where they can reproduce, improve themselves or develop artificial intelligence; that at some point in the next 20 to 50 years, robots could be granted rights. If this happened, the bots would have responsibilities such as voting, the paying of taxes, and serving in the military.
These predictions are contained in papers that look at projected developments over the next 50 years.
If you're thinking we made this up, take a look at http://news.bbc.co.uk/2/hi/technology/6200005.stm.

What's To Come?
Honestly, we haven't the foggiest. But let's take pot-shots at possibilities.

What gets some people's brows all sweaty is questions like "What if they became more intelligent than us?" and "Will they take over?" (That, by the way, has been asked by people about 16 billion times, which is 2.46 times per person)

Turns out, in 2001, someone of no less stature than that of Stephen Hawking recommended that humans change their DNA through genetic modification to stop intelligent machines from "taking over the world." What can we say?

Not much, but Prof Noel Sharkey has license to say something: that contrary to popular belief, robots have no chance of taking over the world. "They don't have as much intelligence as bacteria and they never will." Sharkey is the man behind Robot Wars (see Robo Sapiens, Digit, January 2004), where evolutionary robots were divided into prey and predators, and fought it out. Would Sharkey and Hawking please fight it out and tell us something we can believe!

Perhaps we should believe someone with a name like Marshall Brain. He happens to be the founder of howstuffworks.com, and at the much-read http://marshallbrain.com/robotic-nation.htm, Brain envisions a situation where robots will have all the jobs.

Actually, all talk of "robots on the rise" is complementary to talk about cyborgs and how we will turn cyborg-there's the possibility of a mixed race on the planet, each individual of either machine or human origin, and with sharply varying degrees of intelligence and physical capabilities. Lending credence to this idea is Ray Kurzweil, inventor and futurist: "By the time we succeed in building (intelligent) machines, we will have become part machine ourselves."

Kevin Warwick, the world's first self-proclaimed cyborg, who wrote Human-Machine Convergence for Digit in June 2005, says about the people who don't "upgrade": "If they don't, they could become some sort of subspecies."

The "rise of the robots" doesn't need any spectacular advances in AI, for two reasons: first, AI as it stands today aims at imparting human-like intelligence to machines, and the new machines may not need such intelligence-they could well develop a different kind. And second, even if you do take the stance that our machines are to mimic us, the sheer numbers overwhelm. Take common sense, for example, which is notoriously difficult for a machine to have: with computational power increasing exponentially-and the possibility of quantum computers (see The Quest For Quantumland, Digit, November 2006)-common sense could be hard-coded into artificial brains by brute-force methods, without our even having to understand our own common sense!
As you let your imagination loose and try and imagine the future-which is all we can encourage you to do-try and remember that humanness is overrated. It's gotten down to that.