When third-year students in strategy classes at BI Norwegian Business School have a question about their assignments next semester, odds are a robot will provide their answer. A chatbot—a computer program designed to simulate an intelligent conversation—will respond to routine student inquiries and prompt students to complete assignments, mimicking some of the tasks of a teaching assistant.
Although chatbots are common in other industries like retail and finance, they’ve only recently made their way into higher-education. The idea is that if a chatbot on an e-commerce site can help with common customer-service scenarios, why not bring a chatbot to a course Web site to answer frequently asked questions from students? Some researchers even think the approach might increase engagement, since some students who are embarrassed to ask professors a question in front of their peers might prefer to talk to a software robot.
“Students have a lot of the same questions over and over again. They’re looking for the answers to easy administrative questions, and they have similar type of questions regarding their subjects each year,” says Erik Bøylestad Nilsen, an advisor in edtech and innovation at BI Norwegian Business School. “Chatbots help get rid of some of the noise. Students are able to get to answers as quick as possible and move on.”
Assisting the Teaching Assistants
The chatbot addition to courses at BI next semester is part of a larger project the school is running to test out a new learning management system. Called Differ, the LMS is a product of Edtech Foundry, an Oslo-based tech company that works with universities.
Nilsen says he and his team were searching for a new LMS in 2014 and weren’t satisfied with existing options. They wanted a solution that would not only manage course content, but would “make the students an active part of the teaching and learning process.” They teamed up with Edtech Foundry to pilot Differ, which the company is building with the goal of tackling low student engagement, according to founder and CEO Kristian Collin Berge.
One of the ways Differ aims to spur lackadaisical students to action is by using chatbots to encourage students to interact with course materials. Berge says the bots serve two purposes: First, to respond to frequently asked questions, such as when assignments are due or when to catch office hours. The more forward-thinking use is to have the bots nudge students to take actions like reading a news article or posting to a class discussion forum.
Nina Ronæs, a lecturer in the department of marketing at BI, is used to posting prompts for students in online channels. One exercise she tried was having students identify brands they have a negative associations with and explain why in a class chat online. Berge and his team were interested in how they could automate that process by having a chatbot prompt the students at the time when they’d be most likely to respond.
Building a chatbot doesn’t happen in isolation, Berge says. The machine needs data from courses to learn answers to questions and spot patterns in student behavior. That’s why he’s betting on the marriage of the LMS and these digital teaching assistants. “You can use the channel in which the student and teacher communicate regularly and use that data to train the bot,” he says.
Edtech Foundry worked with BI faculty to manually code triggers that the bot would send to students last semester, similar to the kinds of prompts that Ronæs posted manually. “We see the bot getting sometimes five times higher engagement than when the teacher originally posted,” Berge says, adding that the reasons for success are unclear. It could be because the chatbot sends a direct message to students—rather than posting to the whole group—or because it contacts students at times when are likely to be actively working, instead of the spare moments that professors can grab between lectures to intervene.
Students are also more open to talking to chatbots than Nilsen says he anticipated. They’re opening up about questions they might be ashamed to ask their professor directly, for instance. “They’re afraid of being judged. There’s no space where they can ask the silly questions, where they can stay out of faculty’s loop,” Nilsen says. If a student feels silly asking a question, she might just stay silent and miss out on helpful information. A chatbot could remedy that situation.
New (Inter)faces on Campus
Chatbots made waves at Georgia Institute of Technology earlier this year when a computer science professor used one as a teaching assistant—without telling his students. They were surprised to find out at the end of the semester that Jill Watson, powered by IBM’s cognitive computing technology, had been answering their questions.
Elizabeth Womack, a teaching assistant at Cornell University, says replacing a role like hers with software might help some students, but it hurts graduate students who assist in college courses to gain classroom experience and learn about their teaching styles. She voiced her concerns in a comment to a Wall Street Journal article about Jill Watson: “Having that exposure in such a role has been quite influential in my understanding of the infrastructure required for a course, what works and what doesn't, and how important creating personal relationships with students affects their desire to learn.
IBM and Pearson recently launched a partnership in which Watson will be available to answer students’ questions in some of Pearson’s digital learning materials. Other companies are exploring how chatbots might help students in all aspects of their college careers, from paying tuition to connecting with advisors. CourseQ lets students send inquiries via SMS message to a chatbot that responds to common questions. AdmitHub provides a “virtual campus coach”—a chatbot—that provides text-based support for students 24/7. If a student asks a question AdmitHub doesn’t have the answer to, it connects them with a human who does.
For all the talk about robots replacing human workers, Berge and Nilsen don’t see their chatbot replacing professors or teaching assistants. “It frees up time for the teacher and allows them to focus on high-quality interaction,” Berge says. “It can replace a teaching assistant’s job if the job is only to answer FAQs, but the role can be so much more about facilitating a learning community and following up with students.”