Digital emotions aren’t new. Tamagochi, a children’s craze back in the 90’s, gave us a virtual pet that came with emotions, and responded to stylised pleasure and punishment.
Now engineers are looking to put that concept into the next advance in robotic control.
“Fuzzy logic is great for washing machines, and determinate, adaptive algorithms work for menial tasks. If you look to the animal world, the higher orders of animals are trainable not through direct logic and signals, but through pleasure and pain,” says Doctor Gerard Jung, lead roboticist in Germany’s Klein-Bach Laboratories.
“The beauty of pleasure and pain receptors means that the robot is trained not by a set of pre-calculated goals, rather the various environmental factors, including and especially humans, determine what is right and wrong,” he explains, “This way we let the robot ‘figure out’ what it is meant to do, create its own goals and boundaries. It’s very much like training a small dog or a young child. It’s not quite ‘right and wrong’ in a moralistic sense, it’s physically based at this stage.”