The first spark of empathy has been lit in robots: this is demonstrated by the experiment in which a machine was created capable of predicting the actions of a similar one simply by observing it.
The result, published in Scientific Reports, is due to the group at Columbia University in New York led by Hod Lipson.
It is the first sign of how even in robots there could exist a Theory of the mind, that is, the ability so far considered exclusive to human beings to put themselves in the shoes of others to anticipate their actions.
The result lays the foundations for more efficient communication both between machines and between machines and humans.
The ability to predict future actions on the basis of observation "is a capability that can significantly improve the attunement and the degree of natural interaction of a robot with a human, since the language barrier is overcome in the execution of tasks in which for example, coordination or mutual agreement is necessary "observes Antonio Frisoli, of the Scuola Superiore Sant'Anna in Pisa.
Thinking about future domestic robots, he adds, "we can imagine robots able to support our actions in a collaborative and effective way".
Human beings learn as children to predict the actions of others and this ability is the basis of knowing how to live and work together successfully.
In contrast, robots had not so far been able to implement this type of social communication until the Columbia University experiment.
In the latter a robot had to look for green circles and head towards them, but there was a problem: sometimes the green circle was hidden by a red cardboard box and in that case the robot would either look for another circle green or freeze.
Meanwhile, another robot observed the scene from a perspective that did not hide any green circle and, after observing his partner for two hours, began to anticipate the target and the path of the other machine.
Researchers recognize that the behaviors exhibited by the robot are much simpler than those of humans, however they believe they could constitute a primitive form of empathy.
Also according to Frisoli, the experiment "measured a primitive form of empathy between two robots, understood as the ability of a robot to predict the behavior of another robot in the absence of verbal communication, but only through a visual analysis of the behavior. of the other partner ".
If robots come to anticipate what humans think, important ethical problems may arise.
For example, Frisoli points out, "to what extent must a robot be able to make autonomous decisions based on its prediction? Can a robot, in anticipating man's thought, manipulate man himself and not be more the mere executor of tasks? These are all aspects - observes the expert - that deserve careful reflection of a roboetic and philosophical type ".