Scientists have found that the way people interpret the behavior of anthropomorphic robots and artificial intelligence is related to individual attitudes. Their differences were determined by the neural activity of each person.
The work was coordinated by Agnieszka Wykowska, who studies human interaction with robots. In 2016, her team received funding to study the question of under what conditions people treat them as creatures capable of expressing intentions. In early research, the group was able to prove that humans perceive robots in different ways: some tend to consider their actions deliberate, others - automated. In a new study, scientists at the Italian Institute of Technology have shown that human biases are correlated with different patterns of brain activity.
Researchers tested 52 people. Using EEGs, they recorded the brain activity of the respondents when they were relaxed and not thinking about anything in particular. The participants were then shown the anthropomorphic robot iCub and asked to tell what it was doing. During the experiments, people described the intentions and desires of a machine or automated processes. The results of the experiment are published in the journal Science Robotics.
Measurements have shown that even when the brain is at rest, there is a certain pattern of neural activity in the beta range (associated with cognitive processes and focusing attention). It turned out that he predicts the tendency of people to attribute the ability to act at will to robots. They also found that neuronal activity changed when respondents described the iCub's behavior as mechanistic.
The authors of the work noted that they did not find out whether the observed effect is related to the context in which participants follow the iCub, its appearance, or the general attitude of a particular person towards robots, but they intend to do so in the future.
In addition, according to scientists, bias towards robots can correlate with the emergence of various biases, such as racial and gender. Their new research will be able to show whether certain attitudes towards machines and different social groups are connected at the neuronal level.
“Our findings are striking because they show that bridging the gap between philosophy and neuroscience is possible,” Wykowska said. "People can have different attitudes, such as anthropomorphizing robots to varying degrees, and they can indeed be detected at the neural level." The research findings will be important to understand how humans interact with robots and can predict whether we will be able to take part in our daily lives.