During the game, the robot pretended to be a reasonable creature and people believed him

During the game, the robot pretended to be a reasonable creature and people believed him

The study is published in Science Robotics, where authors modeled the behavior of robots that we might face in the future.

The authors performed a nonverbal Turing test, in which people and a humanoid robot took part, and found that there were specific features of human behavior, in particular the timing of reactions that could be passed on to the robot to behave like a human being.

In Turing's nonverbal test, participants were required to determine whether they interacted with the machine or with a person, and they relied only on how quickly a person or robot pressed the buttons.

The Study Group focused on two main features of human behaviour: time and accuracy in responding to external irritants.

These are characteristics that they had previously compared to get a standard human profile. Researchers used this profile to build their experiment. Participants were asked to respond to visual incentives on the screen. Volunteers played a game with a partner they had not seen.

The results showed that people who interacted with the robot were unable to determine whether it was a person or a robot that was pre-programmed.

This suggests that the robot passed this version of Turing's nonverbal test.