Language Buddy

Contributors: Karim Salem, Evan Bacon, Brittan Robinett

Introduction

The University of Washington (UW) claims on their website that they are “10th in the nation for hosting international students” and “are a leader in global education” (washington.edu/global/international-students/). However, our team of students is not the first to notice or recognize some of the struggles that international students face socially or academically at the university.

The Problem

The problem our team is interested in studying and solving is that ESL (English as a second language) students at UW don’t get many opportunities to improve their English by conversing with others. Anyone trying to learn a language that they don’t speak natively may struggle with aspects such as grammar and sentence structure, vocabulary, or pronunciation. Some of UW’s International and English Language Learning courses don’t grade on grammar and sentence structure, making it hard for students who need help to improve, and UW has recently cut a huge amount of these courses and programs according to one of our research participants. Additionally, the way many people learn vocabulary and pronunciation is through conversing with fluent speakers. Due to COVID, reduced ESL programming available, and some students’ fear of talking to new people who may not understand them fully, many ESL students at UW are isolated even more than before, causing them to miss out on experiences to improve and ultimately negatively affecting other parts of their lives. Our team believes that ESL students shouldn’t have to pay for tutors to get assistance, or need to struggle with making friends, taking general UW courses, or getting the same career opportunities as others after college.

This problem is exactly what led us to start designing for ESL students at UW. We believe an app with functionality to allow students to answer questions and converse with an AI bot instead of an actual person and give tailored feedback will address these issues.

Modes of Research

Our initial research plan was to gather data from students currently or previously enrolled in UW ESL courses via survey, interview, and observation. Through our survey, we were able to gain insight into the experiences of 5 ESL students of Ck He, Vietnamese, Spanish, Korean, and Chinese backgrounds. The survey participants answered questions such as their comfort level with English, duration of time in the US, helpful resources they have used, and greatest pain points that come as a result of not being a native English speaker. In addition to our survey results, we were able to conduct 3 live interviews, 2 of which were with staff for the ESL program at UW, and the last including 2 ESL students who were native Arabic speakers. Each interviewee brought a unique perspective, but we received most helpful feedback from the 2 students and the current ENGL102 teacher who has been involved with UW’s English Language Learning programs for many years. Our third research method, observation, ultimately was not able to be conducted. While we contacted several on-campus cultural organizations, many weren’t holding virtual events at all during that time. Overall, we were happy to receive data from such a diverse group of participants and stakeholders because it helped us understand the key struggles faced by the majority of the population.

Research Results and Themes

During our design research, we interviewed two ESL students to gather detailed input about the problems they were facing and the solutions they believe are best. One common fear we found in those interviews was that a majority of ESL students don’t feel comfortable practicing English with American speakers as they might look down on them for their poor English. This induces anxiety and stress whenever they talk to a better English speaker than them. One suggestion they were interested in was the feature of practicing English with a bot. In that way, they can control the duration and subject of the conversation to practice their English stress-free without any fear of being judged.

Another research finding that helped us design our solution was how our interviewees cited language exchange sessions as one the best resources they can use to develop their English skills. They liked being matched with an individual/group that also wanted to learn about their language/culture such that it is a casual and fun two-sided conversation that is beneficial for both sides, and it doesn’t feel like a “lesson” anymore. This motivated us to consider a design that allows our users to be matched with locals who can teach English and are interested in the user’s native language and culture. However, we ultimately decided not to pursue this solution because there was at least one existing application with the same functionality.

Proposed Design

From our research we found that students desired most was an environment where they could practice having conversations in English without the pressure of being judged for the mistakes they make, the formality of a traditional lesson, and the consequences of the real world. To address these desires we designed a mobile app that would provide users with a bot that they could hold simulated conversations with in various different situations. The bot could walk users through a simulated job interview, or just a casual conversation about how their day was. Once the conversation is over users would get a report back going over how they did, rating them in areas such as pronunciation, grammar, and word choice. Users could also have the option to use a VR headset to further simulate these situations by being able to hold these conversations in a simulated environment. For a job interview this could include having a virtual office room with a virtual Interviewer. Such simulations would add an extra layer of immersion and could help users practice their English abilities under the perceived stress that the real life equivalent of these situations would induce.

Below are a series of sketches that give a rough depiction of what having a conversation with the bot would look like.

Final Thoughts

Our team will continue to iterate on and improve the design described in this article in ways that address the needs of our target group. The next steps will include user testing and digital prototyping of our proposed design.