Are we ready for AI in teaching?

Researchers at University of Central Florida's (UCF) Nicholson School of Communication and Media are seeking to understand the student response to artificial instructors, hoping those findings will help design effective machine learning teaching assistants that can facilitate positive learning experiences.

Findings published in the International Journal of Human-Computer Interaction suggest students will only accept an AI teaching assistant that is effective and easy to talk to.

The use of AI helps ease a teacher's workload including responding to commonly asked questions — a task that can swell enormously over a term or semester. The study cites creation of an AI teaching assistant named Jill Watson, created by a researcher at the Georgia Institute of Technology. Jill was fed thousands of questions and answers commonly asked in the online class that the researcher had taught for many years. With some additional learning and tweaks, Jill was eventually able to answer the students' commonly asked questions accurately without any human assistance.

The UCF study had students read a news article about an AI teaching assistant then surveyed their perceptions of the technology.

Jihyun Kim, an associate professor at Nicholson and lead author of the study, said the discovery that students are most likely to accept an AI-based teaching assistant that is useful and easy to communicate with points to the importance of having an effective AI system.

"I hope our research findings help us find an effective way to incorporate AI agents into education," she said.

"By adopting an AI agent as an assistant for a simple and repetitive task, teachers would be able to spend more time doing things such as meeting with students and developing teaching strategies that will ultimately help student learning in meaningful ways."

To read the full study or download a pdf copy click here.

Image credit: ©stock.adobe.com/au/metamorworks