Could robots really help teachers engage students more fully?
According to Sean Kelly, a sociologist at the University of Pittsburgh School of Education, the answer is yes. Kelly, along with computer scientist Sidney D’Mello, has researched the role of tech in helping teachers ask better, more thought-provoking questions for a number of years.
Rather than relying on one-time test scores to judge learning, Kelly’s work is a new advancement in the field of building systems for day-to-day classroom observation. His focus is on the quality of instruction, specifically, the type of questions teachers pose that prompt cognitive processing and verbal response from students.
Working with three computer scientists and an English Professor, Kelly built a computer model that closely approximates an expert educator’s ability to tell when a teacher was asking students meaningful or “authentic” questions.
“Teacher talk is a critical component of observational tools used to train and supervise teachers,” Kelly noted in an email interview with the Washington Examiner. “Yet, to date, there has been a disconnect between the fine-grained portraits of classroom discourse contained in research studies and what school administrators and instructional leaders, considering the scope of their day-to-day duties, can provide to teachers in real classrooms.”
Teachers often ask about basic plot elements, theme, symbols, and characters, but these questions prompt recall rather than critical thinking.
“Authentic” questions, however, do not have text-level “correct” answers. Instead, their solutions are up for debate and require deeper analysis and support to justify. Students learn more when discussing and defending their ideas; best practice involves coaching teachers to develop these types of questions.
Kelly’s AI system could provide this to real-world teachers as well. Yet, he noted in the interview that, “this technology is not designed for evaluative purposes … D’Mello and his colleagues designed our system specifically for the classroom setting, and so far, teachers have shown an impressive ease of use with the technology. Recognizing and respecting the limitations of that technology is critically important. If we can do that, teachers will appreciate the opportunities for new insights into instruction.”
Currently, the average middle school English class typically has less than two minutes of genuine discussion among students and a teacher, according to Kelly. With AI coaching, this figure could increase dramatically.
Some forms of AI (albeit not in the form of classroom robots) are already helping to simplify the lives of teachers in schools. Technology automates basic activities, such as multiple-choice grading, and also provides teachers with a variety of software programs. Some such programs adapt to students, assess their learning needs, and provide real-time feedback to teachers.
As AI programs develop, they could inspire classroom improvements that previously would not have been possible.
“I am now confident,” Kelly concluded, “we will be able to give teachers helpful feedback with appropriate caution and recognizing when more data is needed.”
Kate Hardiman is a contributor to Red Alert Politics. She is pursuing a master’s in education from Notre Dame University and teaches English and religion at a high school in Chicago.