Foreign language education is growing at an exponential pace around the globe. It’s also an area of learning for which tech companies have poured millions of dollars into creating educational products. One element of language learning that carries a huge amount of potential is the possibility of creating intelligent chatbots that can converse with learners in a new language. Currently, it’s uncertain whether chatbots are sophisticated enough to incite and maintain the interest of students, however.

To examine this question, a study was recently conducted at a university in Japan, where 122 students participated over the course of a 12-week English language class. All students had three conversation sessions each with both a chatbot and a human partner. Students were given questionnaires to assess their levels of interest before and after the different conversation sessions and the course at large.

The researchers found that during the first session, students held a similar level of interest between the chatbot and the human partner. This level of interest significantly fell in conversations with the chatbot during the second and third sessions. For the human partner, however, the level of interest remained the same during all three sessions. This shows that where the human conversation partner maintained students’ interest throughout the 12-week course, the chatbot did not. The researchers think that this could be the case for two reasons. First, there could be a novelty effect for the chatbot, meaning that where students might find it cool to talk to the bot the first time, the excitement wears off over multiple sessions. Second, students might view the chatbot’s language as inauthentic and grow weary of it because of this.

Either way, it’s clear that chatbots need some work before they’re effective enough to be used throughout an entire course. This is particularly true given prior research shows that familiarity more positively impacts learning than novelty. For chatbots to become more familiar with their human conversation partners, they must be able to learn from their interactions, including what the student is interested in and the areas that need the most improvement. This could help the chatbot seem like a more authentic conversation partner, which could help maintain students’ interest. And a more interested student will want to continue learning.

Fyer, L., Ainley, M., Thompson, A., Gibson, A., & Sherlock, Z. (2017). Stimulating and sustaining interest in a language course: An experimental comparison of Chatbot and Human task partners. Computers in Human Behavior, 75(1), 461–468.

Image: via Startup Stock Photos