The group led by engineer Hod Lipson published, in Scientific Reports magazine, the chronicle of an experiment with astonishing results: they managed to build a robot capable of reading the behavior of another machine and predicting its future actions , thus going to create a spark of " artificial empathy ".
Empathy is the ability of man to intuit the actions of another person by placing himself in his shoes and in his state of mind. Scientifically we speak of Theory of the mind , which deals with the ability to understand one's own and other people's mental state. The objective of the study, conducted at the Creative Machine Lab directed by prof. Lipson, is aimed precisely at providing artificial intelligences with this type of social communication in order to make their use more and more efficient. The aim is to obtain robots that can integrate into human society not only in appearance but also in behavior.
The first phase of the experiment saw the placement of a first robot in a box of about half a square meter, inside which it had to move towards any green circle it could see. A second robot , after two hours of observation, was able to predict the goal and the path of his partner with a success rate of 98.5%. Boyuan Chen, professor of computer science and lead author of the study , enthusiastically commented on the result: “Our results begin to demonstrate how robots can see the world from another machine's perspective. The observer's ability to put himself in his partner's shoes and understand whether or not he can see the green circle from his point of view perhaps represents a primitive spark of empathy in robots ”.
The experiment has several limitations, even according to the team, but it represents the first sign that even robots can feel empathy, putting themselves in the shoes of others to anticipate their actions. This means that even in robots a Theory of Mind could manifest itself. However, it is clear that the behaviors exhibited by the first robot are simpler than many actions performed by human beings, therefore also easier to predict. However, the researchers say that further research into the technology that studies the interaction between robots will help scientists develop even more sophisticated robots.
The aim of the study is to obtain Behavior Modeling for robots. It represents an essential cognitive ability at the basis of social interaction between animals, and for this reason we would like to equip robots with it as well. The research group wants to get to the point that the observer robot can predict the behavior of the actor robot only through visual processing, without any previous symbolic information and / or hypotheses on the inputs provided.
Empathy in robots and ethical dilemmas
The efforts of the group of prof. Lipson are part of a broader push to equip AI and robots with a spark of empathy. Euan Matthews ( AI and Innovation at Contact Engine ) argues that for AI to become more empathetic, they will need to be able to consider more than one intention. Humans have more than one, conflicting desires and feelings, and AIs will need to become more flexible when it comes to understanding human intentions .
Antonio Frisoli, full professor of engineering at the Scuola Superiore Sant'Anna in Pisa, was interviewed on the matter. "The ability to predict future actions based on observation is a capability that can significantly improve the attunement and 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, there is a need for coordination or mutual agreement ”. The professor also declared “ today we can imagine robots able to support our actions in a collaborative and effective way”.
Creating empathy for robots as well represents a fundamental turning point in the interaction between machines and between man and machine. However , ethical dilemmas arise , which have already racked the minds of the greatest science fiction writers. "We recognize that robots will not remain passive machines following instructions for long," says Lipson himself, "we hope that policy makers can help keep this type of technology in check, so that we can all benefit from it." As Frisoli says: “Can a robot, in anticipating man's thought, also manipulate man himself and no longer be the mere executor of tasks? These are all aspects that deserve careful ethical and philosophical reflection ”. On the other hand, an initial answer on the evolution of the bond between man and machine was given to us by an AI .
Article by Roberto Guida
The article Turned in the way of robotics: ignited the spark of empathy in robots comes from Tech CuE .