IEET > GlobalDemocracySecurity > Fellows > Mike Treder > Cyber
Building Humanoid Robots
Mike Treder   Feb 21, 2007   Responsible Nanotechnology  

Robots with feelings [will be available] in just 10 years, scientists predicted yesterday. They now claim it is essential to give robots their own emotions if they are to be capable of running independently and efficiently enough to take on a variety of domestic tasks.

More from a fascinating article in the Telegraph (UK):

At present, commercially available robots such as automatic vacuum cleaners are little more than drones capable of carrying out only one task. However, speaking at the American Association for the Advancement of Science in San Francisco yesterday, a panel of robotics experts said robots capable of multiple domestic tasks, that can also provide companionship for their owners, will be available within 10 years. And the scientists claim it is already possible to give robots such “feelings”.

A number of groups around the world are now developing robots that have basic emotions in a bid to motivate the machines.

If a robot feels happy after it has cleaned a dirty carpet particularly well, then it will apparently seek out more dirt to do the same. Similarly, if the robot feels guilt or sadness at having failed at a task, it will try harder next time.

“Emotion plays an important role in guiding attention towards what is important and away from distractions,” said Professor Cynthia Breazeal, one of the world’s leading roboticists based at the Massachusetts Institute of Technology. “It allows the robot to make better decisions, learn more effectively and interact more appropriately.”

How much might be possible with humanoid robots? How close can we get to androids—such as those envisioned by Isaac Asimov—who are virtually indistinguishable from us?

Whatever is achieved, it will take many more small steps. Perhaps, however, a “grand challenge” can accelerate that progress.

This is from a special report in The Daily Yomiuri (Japan):

At times, it may seem as if technology is moving ahead at breakneck speed. But in reality, most technological and scientific innovation saunters forward in a stepwise fashion, building on past success and carefully hedged against serious risk.

Grand challenges are different. By design, grand challenges are dreamed up to push the envelope, to break through barriers, and to ignore limits. Think of projects such as putting a man on the moon or mapping the human genome and you get the idea.

Grand challenges have become a favorite paradigm-shifting mechanism in many scientific and engineering disciplines ranging from mathematics and biology to psychology and astrophysics.

Typically what they all have in common is a blend of imagination and wonder—a sense that something heretofore thought impossible might just be within reach.

In 2002, a group of British researchers set about defining a series of grand challenges. One of those they focused on is to create “a succession of increasingly sophisticated working robots.”

In the true spirit of a grand challenge, the Architecture of Brain & Mind project aims to bring these three disciplines together in a single demonstrable system. Along the way, researchers will have to solve a range of rather thorny problems related to natural intelligence, perception, reasoning, learning, and problem solving.

At first glance, these seem like the standard set of AI problems. However, the project distinguishes itself by including even less understood human capabilities such as curiosity, creativity, and the ability for the system to not only know what it is doing, but also the reasons for its actions (at least in some cases).

One sign of success would be a robot capable of functioning at the level of a 2- to 5-year-old child, which is no small task. Another milestone could be a robot capable of autonomously helping a disabled person around a house without explicit pre-programming about its environment. Either result would be an outstanding achievement by today’s standards.

We, and others, are using a phrase (attributed to Stephen Hawking) so often now that it is beginning to seem trite, and yet never has it been more true: Today’s science fiction is often tomorrow’s science fact.

Nearly all of us living today will witness many things “heretofore thought impossible,” including, perhaps, androids that are eerily close to humans. Other wonders could include radically extended healthy human lifespans, robust artificial intelligence, and, of course, desktop molecular manufacturing.

That last one is most likely to arrive earliest. We expect that the world will have to deal with the transformative and disruptive consequences of advanced nanotechnology no later than 2020, and quite possibly several years sooner.

(Hat tip: KurzweilAI.net)

 

Mike Treder is a former Managing Director of the IEET.



COMMENTS No comments

YOUR COMMENT Login or Register to post a comment.

Next entry: Michael Sandel’s Contribution to the Burgeoning Bioconservative Canon

Previous entry: Release the Robot Monkeys