Kids would rather learn from smart robots than less-smart humans: new study

The impacts of virtual learning on school-age kids is one area researchers have examined in recent years, but now, new research goes a step further, suggesting kids may be more susceptible to the teachings of robots over humans.

A new study published in the Journal of Cognition and Development by researchers from Concordia University has found that kindergarten-aged children prefer to be taught by a competent robot as opposed to an incompetent human - the children's age being an influential factor to this research finding.

“This data tells us that the children will choose to learn from a robot even though they know it is not like them. They know that the robot is mechanical,” said Anna-Elisabeth Baumann, the paper’s lead author and a PhD candidate, in the study press release.

The study researchers observed two groups of Canadian children, one a group of three-year-olds and the other a group of five-years-olds, as they participated in virtual meetings through Zoom.

On the screen, they watched a video of a woman and a small robot, named Nao, who had human-like characteristics such as a head, face and body.

The children watched as the woman labelled objects incorrectly (calling a car a book, for example), while Nao correctly labelled objects.

The next part of the experiment involved the two “teachers” presenting the children with items that they would likely not recognize, including the top of a turkey baster, a roll of twine and a silicone muffin container. This time, the woman and Nao both labelled these unfamiliar items with made up terms like “mido,” “toma,” “fep” and “dax.”

The kids were asked to label the items based on what they just watched. At which point it became evident that the group of five-year-olds were more influenced by Nao than the teacher, as they repeated the robots nonsensical object labels.

“We can see that by age five, children are choosing to learn from a competent teacher over someone who is more familiar to them — even if the competent teacher is a robot,” said Baumann.

On the other hand, the three-year-olds did not seem to prefer the human versus the robot, as they showed no difference in preference for the nonsense words of the objects.

To observe how the robot resembling a human affected the outcome of the study, the researchers swapped in a new robot, this time a small truck-shaped robot called Cozmo.

But the same results occurred as with Nao, suggesting that the makeup of the robot had no significant impact on how children trusted the robot teacher.

Some may wonder whether the kids fully understood that what they were learning from was, in fact, non-human.

But when the kids were asked if the robots were made of biological organs or mechanical gears, the response from the five-year-olds indicated an understanding that robots are mechanically created, suggesting this had no influence on the outcome.

The three-year-olds were confused by the question, and assigned both biological and mechanical internal parts to the robots.

There have been other studies examining similar areas of research – such as looking at the design of school curriculum using artificial Intelligence (AI), or research on how social robots could be implemented in the classroom.

But this study is the first of its kind, comparing a human to a robot in order to observe how children differentiate the two, as well as to determine how trust strategies emerge, according to the researchers.

This research is important, as robots can be used as tools to understand how kids can learn from both humans and non-humans, said Elizabeth Goldamn, Horizon Postdoctoral Fellow and contributor to the study, in the press release.

“As technology use increases, and as children interact with technological devices more, it is important for us to understand how technology can be a tool to help facilitate their learning.”