Imagine a time in the near future when the traditional teacher is no longer human.
Imagine a computer interacting with students in an almost human-like manner.
The Psychology Department at The University of Memphis is one step closer to bringing the idea of computer-driven teaching into reality.
The program they have developed, which has been named AutoTutor, has the ability to carry on conversations with students and is also able to give responses to students’ questions and feedback to their answers.
Currently, The U of M, Massachusetts Institute of Technology and the University of Southern California are the only universities providing the AutoTutor program.
“Students learn about topics by having a conversation with a talking head,” said Arthur Graesser, co-director of the Institute for Intelligent Systems and director of the Center for Applied Psychological Research at The U of M. “AutoTutor asks deep questions to students, and it holds a conversation.”
AutoTutor was introduced to The U of M in 1997 with a grant from the National Science Foundation. It acts as a conventional teacher by asking questions to which students respond by typing in their desired answer. In return, the computer program provides interactive learning, hints and feedback to the students’ answers.
At The U of M, AutoTutor is used in only two courses, conceptual physics and computer literacy.
Not all students at The U of M are sure they would like a computer-simulated teacher for their classes.
Antoris Smith, a senior majoring in history at The U of M, said if given a choice between the AutoTutor program or a human professor, he believes a human professor would be more beneficial.
Smith said when it comes to grading policies, AutoTutor would not be good for students. He said human professors would be able to offer more leeway in their grading policies.
Smith also stressed that interaction with the teacher is essential to classes. He questioned whether the course material would be covered better by AutoTutor or by a human teacher.
Graesser stressed that AutoTutor would not teach entire courses. Rather, it would be used as an “add-on exercise” to course requirements.
“If you compare AutoTutor to just reading text, AutoTutor is better,” Graesser said. “We believe we learn a lot from discourse and conversation.”
Although Graesser said there are no real disadvantages to AutoTutor, one aspect he said the program has some limitations with is humor. Graesser said AutoTutor can be programmed to include humor, but it cannot change with the pace of what is considered humorous from one day to the next.
Also, another obstacle developers encountered is the voice of the program. Graesser said the program tries to emphasize words correctly, and developers used modified human voices for AutoTutor.
“It’s not perfect,” Graesser said. “But it’s good enough. We want to improve it to the point where students feel comfortable having a conversation with the computer.”
