Robots Are Your Friends ... or They Will Be

Nov. 3, 2006 — -- Sporting one of the world's most recognizable faces and a space suit, Albert has met President Bush, and Russian President Vladimir Putin. At 4.5 feet tall and 132 pounds, Albert does tai chi, and he dances.

But Albert is not an astronaut, a politician, a diplomat, a scientist or a gym buff. Albert is a robot, one that could be coming to your home or workplace sooner than you think.

Experts believe that in 2031, robots will be a common part of our daily lives, doing things like teaching, helping with household chores, serving as our companions, and guarding us in dangerous situations.

"One type of general service public robot, whether humanoid or not, will be available at home, much like today, [where] most homes own a computer," said Jun Ho Oh, a professor at the Korea Advanced Institute of Science and Technology at the National University of Korea.

Oh is Albert's father.

The professor wanted his walking robot to have a familiar human face capable of human expression. To accomplish that, he enlisted the help of David Hanson at Hanson Robotics Inc.

Hanson, a renowned artificial intelligence researcher and robot designer, is widely regarded as the creator of the world's most realistic humanoid robots. In this case, he designed the robot's face so it would resemble physicist Albert Einstein's when he was in his 50s.

"The logic behind his choice is that Einstein's face is likely the most recognizable face in the entire world," Oh said. "It is also [a] synonym of peace and intelligence."

The face is made from a spongelike silicone-type material called Frubber. Frubber allows Albert to express a full range of emotions; it can simulate the actions of more than 48 major facial muscles.

"Frubber material allows for a range of facial expressions, from wide-eyed surprise to angry frown[s] and from smiling to looking sad," Hanson said. "It can even express the subtle difference between intense concentration to really intense concentration."

Hanson put cameras in Albert's eyes, and powerful artificial intelligence software, which helps Albert see and track human faces and moves, understand speech and hold realistic conversations. The robot can even recognize people's faces. If it doesn't recognize your face, it may ask you your name.

Oh sees privacy as a major advantage in the age of robots.

"Who heard of a robot maid writing a tell-all book?" he joked. "It should be a hit with celebrities!"

Oh believes privacy is extremely important for those who need help moving around and performing basic human tasks.

"The robot would be able to bathe people, help them dress, feed them ... without making people feel they have lost their privacy and dignity," he said. "If you have a robot, he or she won't complain."

For now, Albert is a passive robot that acts on commands and serious programming. Both Oh and Hanson agree that in 25 years, robots will be proactive and able to read their owners' intentions.

However, they warn that, along with more independent "thinking," there are safety, liability and legal issues.

"A robot ... could read the owner's intentions wrong," Oh said. "An unstable robot could cause involuntary harm, and these are issues that still need resolve."

The men disagree over the possibility of a "Terminator" scenario, where machines would take over humankind.

"That's impossible, because artificial intelligence is computer intelligence and although it can imitate intelligence, it cannot be intelligent," Oh said. "Think of [Rodin's] "The Thinker" statue. We think he is thinking, yet he is just positioned in that way. [With robots,] there is no thinking involved."

Oh believes the only potential danger for a robot's instability would be caused by the inability of programmers to resolve a complex program crash that could cause a robot to act up.

"It would be a giant error," he said.

Hanson, on the other hand, believes that with computers getting smarter every two years, the "Terminator" scenario would be quite possible unless scientists teach artificial intelligence compassion and how to read human emotions.

"Emotions are helpful for a social species, but dispensable for an entity that does not need the assistance of other entities in order to survive ... or reproduce," Hanson said. "Without compassion and human emotion, robots could become sociopaths."

When it comes to being social, Grace, a robot from Carnegie Mellon University, takes the title. Grace is a socially interactive mobile robot. She was an instant hit when she attended a conference in Canada in 2003.

"She is a six-foot-tall, drum-shaped robot whose head is a computer monitor with a highly expressive, digitally animated female face," said her handler, Greg D. Armstrong.

At the conference, Grace had to find her way into the convention center, register for the conference, take an elevator and enter the room where she was to deliver a speech and then deliver the speech all by herself. She managed to do it all without any assistance.

Grace is used in research on social interaction with people while moving around human environments.

"She uses a laser and a camera to navigate through rooms, around furniture, and with people, trying to figure out people's intentions so that she can move in a socially appropriate manner," said Marek Michalowski, a Ph.D. student at Carnegie Mellon who develops and studies human-robot interaction with Grace.

"Meanwhile, she uses her personality to ask for help, entertain people, and provide assistance to those she interacts with," he said.

Similarly, Quasi, an animatronic robot created at Carnegie Mellon's Entertainment Technology Center, is used to study how to design characters that are emotionally expressive, approachable and engaging.

"Quasi has a number of features that allow him to express emotion and personality, including color-changing eyes and antennae, upper and lower eyelids that behave like eyebrows, and antennae that can move forward and backward, as well as in and out, giving them an expressive quality -- not unlike that of a dog's ears," said Seema Patel, a Carnegie Mellon University graduate and CEO of Interbots.

The Carnegie Mellon team reported that most people who meet Quasi quickly form a strong emotional connection with him.

Twenty-five years from now, the same human-computer interaction design principles that allow Quasi to form such strong bonds with people, will be vital when creating these next-generation interactive devices, according to Patel.

As devices and robots become more and more interactive, people will expect them to exhibit believable personality and emotion. It will be important for engineers to design products capable of meeting consumers' expectations and forming connections with their owners.

"I wouldn't be surprised if, in 2031, your microwave, television , and computer were all just as friendly and expressive as Quasi is," Patel said.