Researchers at the UCSD Machine Perception Laboratory published a study that suggests that babies time their smiles to elicit a smile in return. The researchers detailed their findings in the Sept. 23 issue of PLOS ONE, the study being a part of an effort to better understand human development through the use of robots.
Headed by Javier Movellan, a research scientist at the Machine Perception Laboratory and an alumnus of the Jacobs School of Engineering, the team conducted its study by designing and constructing Diego-san, a robot that mimics toddler-like behavior.
The researchers programmed Diego-san to have four different types of behavior in order to emulate the babies they have studied. One such set of behavior was for Diego-san to smile back every time the person it interacted with smiled.
In testing their theory, Diego-san interacted with 32 UCSD undergraduates for individual three minute sessions. Researchers observed similarities between these interactions and communications with human infants; Diego-san smiled as infrequently as possible, yet still caused the undergraduates to smile quite often.
During the experiment, Diego-san confirmed the team’s findings. By using data from another study that observed the interactions between 13 infants and their mothers and running that data through an algorithm, the team found that the act of smiling was not sporadic or random for babies. Instead, 11 of the 13 babies in the study clearly showed intentional smiling.
Dr. Leslie Carver, a professor in the department of psychology at UCSD, concurred with Ruvolo and said that babies are active participants in social interactions and that smiling cannot be attributed to random behavior.
“When they first start smiling, babies aren’t just behaving randomly — they’re systematic and doing something that has some sort of intent behind it,” Carver told the UCSD Guardian. “I’m not sure that you could say babies are consciously aware of what they’re doing, but their behavior is predictable and not just random.”
Carver also mentioned that people should not jump to conclusions based on the actions of a robot.
“I would caution people against assuming that because a robot can be made to produce the same behavior as an infant, that the infant is necessarily doing the same thing as the robot,” Carver said.
While stating that Diego-san supports the researchers’ hypothesis that babies act in certain ways intentionally, Carver thinks that there could be other reasons at play.
“It’s possible … that babies have a ‘goal’ when interacting with an adult — but there might be other reasons that the robot and the baby would behave in similar ways as well,” Carver said.
In addition, the creation of Diego-san and the researchers’ findings can help us better understand human development.
“We still don’t know very much about how infants develop social connections with other people. The research tests an important model of how this process begins,” Carver told the Guardian. “I think that the results show that there is an important role for socially rewarding experiences in how infants develop sociability, and I’d be interested in seeing research that extends this to later ages and longitudinal studies.”