One more stepping stone...

[Arkan]

To the robot take over.

 

RICHARDSON, Texas - David Hanson has two little Zenos to care for these days. There's his 18-month-old son Zeno, who prattles and smiles as he bounds through his father's cramped office. Then there's the robotic Zeno. It can't speak or walk yet, but has blinking eyes that can track people and a face that captivates with a range of expressions.

 

At 17 inches tall and 6 pounds, the artificial Zeno is the culmination of five years of work by Hanson and a small group of engineers, designers and programmers at his company, Hanson Robotics. They believe there's an emerging business in the design and sale of lifelike robotic companions, or social robots. And they'll be showing off the robot boy to students in grades 3-12 at the Wired NextFest technology conference Thursday in Los Angeles.

 

Unlike clearly artificial robotic toys, Hanson says he envisions Zeno as an interactive learning companion, a synthetic pal who can engage in conversation and convey human emotion through a face made of a skin-like, patented material Hanson calls frubber.

 

"It's a representation of robotics as a character animation medium, one that is intelligent," Hanson beams. "It sees you and recognizes your face. It learns your name and can build a relationship with you."

 

It's no coincidence if the whole concept sounds like a science-fiction movie.

 

Hanson said he was inspired by, and is aiming for, the same sort of realism found in the book "Supertoys Last All Summer Long," by Brian Aldiss. Aldiss' story of troubled robot boy David and his quest for the love of his flesh-and-blood parents was the source material for Steven Spielberg's film "Artificial Intelligence: AI."

 

He plans to make little Zenos available to consumers within the next three years for $200 to $300.

 

Until then, Hanson, 37, makes a living selling and renting pricey, lifelike robotic heads. His company offers models that look like Albert Einstein, a pirate and a rocker, complete with spiky hair and sunglasses. They cost tens of thousands of dollars and can be customized to look like anyone, Hanson said.

 

The company, which has yet to break even, was also buoyed by a $1.5 million grant from the Texas Emerging Technology Fund last October. The fund was created by Gov. Rick Perry in 2005 to improve research at Texas universities and help startup technology companies get off the ground.

 

Hanson concedes it's going to be at least 15 years before robot builders can approach anything like what seems to be possible in movies. Zeno the robot remains a prototype.

 

During a recent demonstration, Zeno could barely stand and had to be tethered to a bank of PCs that told it how to smile, frown, act surprised or wrinkle its nose in anger.

 

Robotics, Hanson believes, should be about artistic expression, a creative medium akin to sculpting or painting. But convincing people that robots should look like people instead of, well, robots, remains a challenge that robot experts call the "uncanny valley" theory.

 

The theory posits that humans have a positive psychological reaction to robots that look somewhat like humans, but that robots made to look very realistic end up seeming grotesque instead of comforting.

 

"Nobody complains that Bernini's sculptures are too darn real, right? Or that Norman Rockwell's paintings are too creepy," Hanson said. "Well, robots can seem real and be loved too. We're trying to make a new art medium out of robotics."

 

So just how did Hanson end up with two Zenos, anyway?

 

It all goes back to when his wife, Amanda, gave birth to their first child and Zeno the robot was already in the works.

 

They rattled off several names to their baby boy, but it wasn't until they whispered "Zeno" that "this look of peace fell over his face; it was like soothing to his ears," Hanson recalled.

 

"There was no way we could give him any other name. He chose Zeno as his name," he said.

 

That was just fine with Amanda.

 

"I thought that it was very endearing, very sweet," she said.

 

The similarities go beyond the name. Though Zeno the robot was built to resemble the animated Japanese TV show character Astro Boy, his plastic hair and saucer-shaped eyes bear a striking resemblance to the curly locks and wide-eyed smile of the real Zeno.

 

"So by coincidence they're both Zeno, and in other ways this robot has become more of a portrait sculpturally of the son, although it's almost coincidence," said Hanson, whose previous jobs include working as a character sculptor for The Walt Disney Co. "We didn't consciously sculpt this robot to look like him. It's the way things filter through the hands of the artist."

 

Hanson says one of the robot Zeno's biggest advancements is that its brains aren't inside the robot. Instead Zeno synchs wirelessly to a PC running a variant of Massive Software

[Aram]

I bet it doesn't work worth a ****.

[Dark_Raven]

Do Androids Dream of Electric Sheep?

[Walsingham]

Roll on the Butlerian jihad!

[Musopticon?]

I could do with mentat training already. Ancient history lectures are a bitch to take notes of.

[Pidesco]

Taking notes in class is for the weak.

[Musopticon?]

Well, we do have an interactive wiki enviroment for all lecture material, so I agree, but I don't yet own a computer.

[Azarkon]

People outside of AI and Robotics really can't grasp just how far away science actually is from achieving their (rather unreasonable, once you think about it) expectations. That said, it's common sense that "presentation is everything." If you can get people to think that they're interacting with an intelligent being (ie like with those early chatbots), it really doesn't matter whether the being is actually intelligent or not. There's where I expect a great deal of progress in the coming years.

Edited September 16, 2007 by Azarkon

[Arkan]

People outside of AI and Robotics really can't grasp just how far away science actually is from achieving their (rather unreasonable, once you think about it) expectations. That said, it's common sense that "presentation is everything." If you can get people to think that they're interacting with an intelligent being (ie like with those early chatbots), it really doesn't matter whether the being is actually intelligent or not. There's where I expect a great deal of progress in the coming years.

 

Nobody knows for sure what will be possible in the future. Some "breakthrough" could take place that could "revolutionize" AI and all that rot.

[Fenghuang]

Nanotech powered microprocessors should be able to give future AIs a lot more power assuming we're able to code them.

[Walsingham]

I could do with mentat training already. Ancient history lectures are a bitch to take notes of.

 

I must not attend class. Classes are the mind killer.

[Musopticon?]

Classes are the little-death.

[Guest The Architect]

Taking notes in class is for the weak.

 

Listening to Pidesco is for the weak.

 

Sorry, I have nothing against you, I just like to randomly insult people.

[Pidesco]

Classes are the little-death.

 

Classes are orgasms? I can I see why the Finns have a successful educational system.

 

Taking notes in class is for the weak.

 

Listening to Pidesco is for the weak.

 

Sorry, I have nothing against you, I just like to randomly insult people.

 

Don't worry. It's true, anyway.

[Azarkon]

Nobody knows for sure what will be possible in the future. Some "breakthrough" could take place that could "revolutionize" AI and all that rot.

 

Certainly, but I believe a level of technological sophistication is required before breakthroughs can occur, and current research across the sciences simply hasn't reached that stage. Practical high-performance computing via nano-technology is a ways off, and quantum computing remains a fantasy. Neuroscience is not much closer to understanding how the brain works beyond primitive reactions and psychology is still a dead-end. Maybe biological computing could do something, but building a brain is not exactly the same as building AI. We could already accomplish the former (via cloning, for example, if the technology wasn't banned), but it gets us no closer to understanding intelligence, which is a prerequisite of translating it into computers.