Humans behave and act in ways that other humans can recognize as human. If humanity has specific characteristics, is it possible to reproduce these characteristics on a machine like a robot? Researchers from the IIT-Istituto Italiano di Tecnologia (Italian Institute of Technology) tried to answer this question by implementing a non-verbal Turing test in a human-robot interaction task. They involved human participants and the humanoid robot iCub in a joint action experiment. What they discovered is that specific characteristics of human behavior, namely response time, can be translated into the robot in such a way that humans cannot distinguish whether they are interacting with a person. or a machine.
The study was published in Scientific robotics newspaper today and is the first step to understand the kind of behavior that robots could exhibit in the future, considering the different possible fields of application, such as healthcare or the manufacturing production chain.
The research group is coordinated by Agnieszka Wykowska, head of the “Social Cognition in Human-Robot Interaction” laboratory at IIT in Genoa, and beneficiary of the European Research Council (ERC) for the project called “InStance”, which addresses the question of when and under what conditions people treat robots as intentional agents.
“The most exciting result of our study is that the human brain is sensitive to extremely subtle behavior that manifests humanity” – comments Agnieszka Wykowska. “In our non-verbal Turing test, human participants had to judge whether they were interacting with a machine or a person, considering only when buttons were pressed during a joint action task.”
The research group focused on two fundamental characteristics of human behavior: time and precision when responding to external stimuli, characteristics they had previously mapped to obtain an average human profile. The researchers used this profile to build their experiment where participants were asked to respond to visual stimuli on a screen. Participants played the game split into two human-robot pairs: one person teamed up with a robot, whose response was controlled by the other pair’s person or in a pre-programmed way.
“In our experiment, we pre-programmed the robot by slightly varying the average human response profile,” says Francesca Ciardo, first author of the study and Marie Sklodowska-Curie Fellow at the Wykowska Group in Genoa. “In this way, the possible responses of the robot were of two types: on the one hand, it was fully human since it was controlled by a remote person, on the other hand, it lacked certain human functionalities because it was pre -program “.
The results showed that people interacting with the robot were unable to tell whether the robot was human-controlled or pre-programmed in the state where the robot was actually pre-programmed. This suggests that the robot passed this version of the non-verbal Turing test in this specific task.
“The next steps in scientific investigation would be to engineer more complex behavior on the robot, to have more elaborate interaction with humans, and to see which specific parameters of that interaction are perceived as human or mechanical,” concludes Wykowska.