A New Face For Robotics

tanmay writes "Android technology has moved a step forward with the creation of a high-tech polymer called 'f'rubber,' which resembles human skin. Its creator, David Hanson has implemented it in a robot called Hertz, as this report from CNN gives us the details. Another question that the report brings up is the need to make robots resemble humans. Ray Kurzweil thinks Hanson's work is significant because realistic facial movement will play an important role in the way future androids respond to humans, and has the following to say, 'Intelligence significantly below that of normal humans stands out more with a robot that looks strikingly human. This creates the impression of a human with impaired intelligence, which may strike some as disturbing.'"

Comments

New Scientist recently covered this as well

[another misanthrope]

full text here: Follow that human


"Most people doing social robots believe that human faces will turn people off and will disturb them. I think that's ridiculous," Hanson said. "The human face is perhaps the most natural paradigm for us to interact with."

Most experts disagree. They cite one of the principles of social robotics, the so-called "Uncanny Valley" theory.

First described by pioneering Japanese roboticist Masahiro Mori, the theory goes like this: humans have a positive psychological reaction to robots that look somewhat like humans. But if a robot is made to look very realistic but somehow isn't quite right (it has an odd smile, or it doesn't blink, for example) it seems grotesque instead of comforting.

PopSci article

[FiloEleven]

Popular Science has a longer (and IMO better) article on the entire project. It was written September 2003. It's got interesting information on the "Uncanny Valley" -- robots are okay, unless they look very much but not quite human - they call it "walking corpse." Hanson hopes to get past that valley and build (at least) a head that is a perfect human imitation.

Heard Ray speak last week

[tundog]

He gave 2-hour talk about the relationship between innovation, AI and biotech.

The coolest part was that his talk was a virtual talk - he was sitting at his office in Boston and was beamed over to a conference hall with ~2000 people. They had this curtain setup with a translucent concave reception dish that caught a projected video signal - I swear to god, from the back of the room, the only way that you knew he was a hologram and not a real person was that he was 'brighter' than the guy next to him. Even better, was that they had this camera that projected the people speaking onto a huge screen auditorium-type display and when you looked at that there was no way to tell that he wasn't physically there. The only thing that gave him away was the occasionally interrupted audio (must have been VOIP). I don't know if the video signal was analog or digital but I suppose it could have been either.

The core of his talk was that science in general (and machine AI in particular) is advancing 'exponentially' - that each new innovation provides us with new tools to accelerate progress. Cool shizzle. According to him, we'll see some incredible advances in the next 10 years.