(April 8, 2010) -- The University of Tokyo's JSK Robotics Laboratory recently unveiled Kojiro, a new humanoid robot designed to operate in a household environment. Built from lightweight, flexible material on a skeletal structure, the robot's appearance and movements are strikingly human. But are robots such as these really alive?
According to Idaho National Laboratory roboticist Derek Wadsworth, you have to first differentiate between mere automatons and actual robots to weigh the question properly. An automaton -- such as a mechanical manipulator on a vehicle assembly line -- has the ability to sense and act. For example, as a vehicle frame advances on the conveyer belt, the manipulator senses its presence and then installs a windshield.
A true robot applies one more step to the process: reason. It analyzes the sense data and then acts based on its computations.
"As humans, we have those same capabilities," Wadsworth told Discovery News. "In everything we do, we sense our environment, we analyze what we're sensing and then we produce an action. So are robots alive? They're alive in the sense that they have the sense, reason and act capability."
Where robots fall short, however, is that they require humans to tell them what exactly they need to sense in order to act.
"I don't believe I myself understand everything that I sense in my environment," Wadsworth said. "It just kind of automatically happens. There are some obvious things we can sense -- touch, smell, heat, light, sound -- but there are a lot of ancillary sensory perceptions based on our life experiences that we reason and act on."
This is where the study of artificial intelligence really comes into its own: programming robots with the ability to navigate a sea of sense data and determine what should influence behavior and decision making. If you think about it this way, robotic household vacuum cleaners already exhibit crude learning.
"A Roomba goes around bouncing off the wall and pretty soon it begins to understand, 'Hey, I'm expecting a wall in this room,'" Wadsworth said. "So it avoids that wall, and it eventually forms a map of the room so it doesn't need to bounce off the walls anymore."
Still, the question of life continues to divide even the robotics world, according to Dr. Ayanna Howard, director of Georgia Tech's Human-Automation Systems Lab.
"Some say robots are alive because they can have emergent intelligence," Howard said. "They can emerge behaviors and therefore, if they're thinking, acting and learning, then they have semblance of being alive, just like people. And then you have more the philosophical world that says no, our consciousness and sense of being is what makes us alive, and while robots that can be programmed to have a sense of being, they themselves won't awaken to it."
Given current technological trends, Howard says that improved artificial intelligence and robotic/biological hybrids will complicate the question further. Human perceptions also continue to skew any attempt to sort the appearance of life from life itself.
"As a robotics engineer, I don't think robots are alive because I program them," Howard said. "But if I give the robot to my son and endow it with the capabilities to seem autonomous and exhibit emotions, then my son will believe most likely that it is alive. And that is actually my goal: to create robots that seem to be cognizant, that seem to be intelligent."