by How would you feel if your child was taught by a robot?
Social robots – robots that can speak and mimic human emotion – have been introduced into classrooms around the world. Researchers have used them to read stories to preschoolers in Singapore, to help 12-year-olds in Iran learn English, to improve handwriting among young children in Switzerland and to help students with autism in England to teach appropriate physical distance during social interactions.
Some experts believe that these robots could become “as common as paper, whiteboards and tablet computers” in schools.
Because social robots have a body, people react to them differently than we do to a computer screen. Studies have shown that young children sometimes accept social robots as peers. For example, in the handwriting study, a 5-year-old boy continued to send letters to the robot months after the interactions ended.
As an education professor, I study the different ways teachers around the world do their jobs. To understand how social robots might affect teaching, graduate student Raisa Gray and I introduced a 4-foot-tall humanoid robot called “Pepper” into a public elementary and middle school in the U.S. Reveal our research many problems with the current generation of social robots. , making it unlikely that social robots will be running classrooms anytime soon.
It’s not ready for prime time
Much of the research on social robots in schools is done in very limited ways. Children and social robots are not allowed to interact freely with each other without the help, or intervention, of researchers. Only a few studies have used social robots in real-life classroom settings.
Also, robotics researchers often use “Wizard of Oz” techniques in classroom settings. That means a person is operating the robot remotely, which implies that the robot can actually talk to people.
Limited social skills
Robots must be quiet.
Background noise of any kind – class change bells, loudspeaker announcements or other conversations – can interfere with the bot’s ability to follow a conversation. This is one of the major problems facing the integration of robots in schools.
It is extremely difficult for programmers to create software and hardware systems that can achieve what humans do unconsciously. For example, the current generation of social robots cannot interact with a small group and, for example, track the facial expressions of multiple people. If a person is talking to two other people about their favorite football team and one of the listeners is crying or rolling their eyes, it’s likely that someone will get it.
A robot will not.
Also, unless a barcode or other identification device is used, today’s social robots cannot identify individuals. So they are less likely to have realistic social interactions. Facial recognition software is difficult to use in a room full of moving and shifting people, and it also raises serious ethical questions about keeping students’ personal information safe.
The dialogue is pre-programmed
In order to operate the robot, our students had to master the tutorials that came with the robot. Some students quickly realized that the robot could only respond to certain basic routines.
For example, Pepper could answer “How old are you?” but not “How old are you?” Other students still wanted to interact with the robot as if it were a human and became very frustrated with the non-human responses.
When a robot fails to answer a question, or answers in the wrong way, students realize that the robot doesn’t really understand them and that the robot’s dialogue is pre-programmed. The robot cannot make sense of the social context.
In our study, the students learned to adapt to the robot.
A group of girls would stand around the robot and one of them kept hitting his head. This caused the robot to perform its “I feel like a cat” or “I’m ticklish today” routine. The girls seemed to like this. They seemed happy that one person was interacting with the robot while others were watching.
It is not possible to move around the classroom with ease
Students who have seen YouTube videos of robotic dogs running and jumping may be disappointed that most social robots cannot move around a classroom with ease. The teachers in our study were disappointed that Pepper could not bring them coffee.
These problems are not limited to school settings.
Some healthcare facilities have service robots programmed to deliver medicine, but this requires special sensors and programming. And while stores and restaurants are experimenting with delivery and cleaning robots, when a grocery store in Scotland tried to use Pepper for customer interactions, the robot was canceled after a week.
What social robots can teach children
Although the social robots currently used in schools are limited and limited in functionality, they can still provide useful learning experiences. Students can use them to learn more about robotics, artificial intelligence and the complexity of real human behavior.
As one researcher wrote, “Robots act as a bridge that enables students to understand people.”
Struggling with the limits of robots gives students real insight into the complex nature of human social interaction. The opportunity to work hands-on with a social robot shows students how difficult it is to program robots to mimic human behavior.
Social robots can also provide students with important learning opportunities about artificial intelligence. In Japan, Pepper is being used to introduce students to generative AI. Students can connect ChatGPT with Pepper’s physical presence to see how much AI improves Pepper’s communication and makes it more lifelike.
As AI becomes a larger part of our work and our lives, educators need to prepare students to think critically about what it means to live and work with social machines. And with the real guidance and supervision of a human teacher, students can explore why we want to talk to robots as if they were people.
This article is republished from The Conversation, a non-profit, independent news organization that brings you facts and analysis to help you make sense of our complex world.
It was written by: Gerald K. LeTendre, Penn State.
Read more:
Gerald K. LeTendre receives funding from the Harry L. Batschelet II Endowed Chair within the College of Education, Pennsylvania State University
