In Robots in the classroom: Differences in students’ perceptions of credibility and learning between “teacher as robot” and “robot as teacher”, Edwards presents a study in which college students perspectives of robots in the classroom setting was tested, and argues that students are more likely to follow suggestions made by an autonomous social robot.
Edwards defines social robots as autonomous and able to follow behaviors associated with its role, and explains some of the benefits robots have shown previously in the classroom including improving math performance, increasing retention rates and reinforcing social behaviors in autistic children. Edwards also explains that the success of robot use in the college setting will depend on how credible the students view the robots to be, and how likely they think they will be able to learn from them. In the study, 86 students were placed into robot as teacher (social robot) and teacher as robot (teacher connected through robot) treatment groups. The participants were exposed to 5 minute lectures, and students were led to believe that both bots were autonomous. The results of the study showed that the telepresence instructor as well as the robot teacher were viewed as credible, however the telepresence robot was rated as more credible than the autonomous social robot. Edwards argues that this result is due to social presence theory, which is the idea that someone is communicating with a social entity rather than a inanimate object. Interestingly, participants reported greater likelihood to take behavioral recommendations from a machine instructor versus a telepresence human instructor, which is contradictory to the original findings. Edwards justifies this by saying that telepresence may be better for creating positive attitudes about a subject matter, while the social robot may be more effective at behavior modifications.
Overall, I found Edwards’ argument valid overall and interesting, but I do see many unanswered questions. One thing that I wonder about is whether the subject matter/content plays a role in the types of robots students would rather interact with. Is it more difficult to learn math or coding from a robot teacher?