Future-proofing community interpreting volunteerism through design: a case of bbb Korea
Can human interpretation survive in an artificial intelligence jungle?
Have you ever found yourself needing to talk to people and you don’t speak their languages? Frustrating, to say the least. You just wish you were with someone who speaks both languages and can help you communicate.
Many of us have experienced language barriers at some point in our lives. In Korea, with the dramatic increase of foreign residents/visitors over the recent decades, more and more people — both Koreans and internationals — are experiencing language barriers ranging from a simple conversation in a supermarket or a restaurant to an emergency situations at a police station or a hospital. And this is the challenge bbb Korea has been working to overcome through its phone-based community interpreting volunteerism over the past two decades.
bbb Korea is a non-profit organization that started in 2002 when a large number of internationals suddently started visiting South Korea for the 2002 FIFA World CUP Korea/Japan. Ever since, bbb Korea has been helping people with language barriers in Korea by providing a free 24/7 phone interpretation service between Korean and 20 other languages through volunteerism. Now, there are about 4,500 registered volunteers and over 1,700 daily-active ones with bbb Korea who pick up their phones and interpret for people on the other side of the line breaking down language barriers.
However, technology continues to develop, and thanks to artificial intelligence, online machine translation services are only getting better— they’re easy to use, abundantly available often free of charge, and increasingly accurate. This makes it easy to imagine a future where bbb Korea makes less and less impact on the society. Facing these challenges, bbb Korea approached New Design Studio to think about what bbb Korea should do in order to stay relevant in the ever-changing technological world.
By adapting and utilizing technology — rather than competing with it — our first goal was to make bbb Korea’s services easier for both the requesters and the volunteers, while making it possible to process the large amounts of voice data bbb Korea is collecting, so that it can prepare for the future. Our second goal was to envision a novel service/business model for bbb Korea to prepare for the world where machine interpretation becomes as accurate as community interpreters are today. In this article, we focus on our work in improving the current service and enabling data processing.
Listening in and service safari
We started this project by listening to 200 recorded phone calls provided by bbb Korea — 150 in English, 50 in Indonesian, and 50 in Russian. At the same time, we conducted service safari where some of us experienced bbb Korea’s interpretation service as requesters, and the rest registered as bbb Korea’s volunteers. As requesters, we requested for interpretation in English, Indonesian and Russian and made interpretation requests from a restaurant, pharmacy, camping site, and mobile phone store. As volunteers, we received interpretation requests from a hospital, a mobile phone store, a police station, an online store and more.
Listening in gave us the initial understanding of how the service and volunteerism works. We could already tell how bbb Korea helps break language barrier in various situations — helping people on both sides beyond mere interpreting. Many of the calls we listened to were about the questions about COVID-19 quarantine. Sometimes the public servants did not understand what the requester was not aware of, and the interpreters proactively gave the information they gained through multiple interpretations to the requester, or asked questions the requesters did not ask to the public servants to bridge the gap. This proactive role beyond literal interpreting was a good surprise and convinced us that bbb Korea’s volunteerism will stay relevant for quite a while from today.
Experiencing the service firsthand gave us a better understanding of how the service works for both volunteers and users, as well as their needs, gaps, and opportunities. One of the calls we processed helped the requester talk to a police officer after a car accident, and another helped the requester schedule a meeting with mechanic from an insurance company after the car battery ran out. With these experiences, we could more easily empathize with the users of bbb Korea’s service — both the requestors and volunteers.
Based on the basic user study above, we gained insights about the need for providing contextual information of the requesters to the volunteers and adding a simple chat feature to solve some sticky problems.
For instance, the recorded phone calls revealed that a large number of requesters forget to introduce themselves, where they’re calling from and why they need the help. This lack of contextual information makes it hard for the interpreter to get a grasp of the situation, which led to a situation where they spend minutes to figure out the purpose of the call. So, we ideated the ways in which requesters could input their location and situation using hashtag-like buttons before making the calles.
Further, some types of information are very difficult to verbally deliver over the phone from Korean to other languages and vice versa, such as addresses and names of people. A simple chatting feature that lets the interpreter and requester communicate through text and photos would help solve this problem! These actions could enable labelling the recording files made in the future, enabling augmented data analysis by A.I.!
Concurrently, we also had extensive discussions with our collaborator Expressive Computing Lab in UNIST in order to identify opportunities for using machine learning and big data to improve bbb Korea’s services. For example, while we were listening to 200 interpretation recordings, EXPC used A.I. to transcribe thousands of recordings, extracted keywords from them and see the occurences of certain keywords in the discussions. As there are at least two languages and often three people talking in one call, it was not easy for the current technology to identify the theme of the discussion. Our ideas for contextual information could help with that!
Staff workshop and follow-up interviews
These ideas were good, but they were still raw, and only improving the app itself was obviously not enough for this project. We still wanted to know what challenges — both internal and external — bbb Korea is facing right now and will face in the future, as well as their hopes and vision. We visualized our ideas into visual prototypes and brought them to our first online co-design workshop with bbb Korea staffs. The reason we turned our ideas into screens was not necessarily to test and validate them — rather, it was to to spark conversation and break down bbb Korea’s fixation on their own service by exposing them to possible new directions.
The 2,5 hours we spent together on Zoom and Miro board were not enough for us to talk about everything, so we conducted follow-up interviews with the bbb Korea’s staff to listen to their stories and opinions. We had good conversations with bbb Korea’s staff about their current and future challenges, opportunities, and visions in the next decades.
Early-stage ideation
To make sense of the insights we got from what we’ve done so far — social listening, listening to recorded phone calls, service safari, staff workshop, interviews with bbb staff — we conducted two rounds of data analysis.
The first round generated over 30 groups of insights, but here are some of the key insights that would highly influence our second round of affinity diagramming:
- The quality of interpretation provided by bbb is not always consistent, but it’s difficult giving direct feedback to volunteers considering they are doing it out of goodwill. bbb Korea needs to find a way to encourage their volunteers to improve their interpretation and communication skills without anyone getting their feelings hurt.
- Most minority languages spoken in Korea volunteers are exhausted because they get a lot of bbb calls and for them its hard to provide fun elements about volunteering.
- bbb Korea wants to go beyond interpretation — aiming to contribute to building a more integrated and inclusive society
Based on our first round of analysis, we proceeded to do a second round which focused more on generating new ideas for bbb Korea. The ideas revolved around these how-might-we questions:
- How might bbb Korea empower its volunteers to improve the service themselves with ownership?
- How might we create protective boundaries for bbb volunteers?
- How might bbb Korea empower/support its volunteers to help build a more culturally inclusive society?
- How might we help those who are living in, or visiting, Korea but do not speak Korean?
Digging deeper–stakeholder and user interviews
We considered the ideas generated during above stages immature and wanted to develop them further. We also knew at this point that we need to dig deeper into the problems, needs, and opportunities of stakeholders and various users to ideate further. Therefore, we conducted interviews with 2 bbb Korea stakeholders, 6 potential users of bbb Korea’s services, and 6 interpreters including those who volunteered for bbb Korea.
After spending quite some time doing interviews, analyzing the insights, and creating more ideas, we came up with what we called ‘idea packages’. Not all of these idea packages were intended to be developed further to be an actual design solution, but to work as anchors to discuss further with bbb Korea’s staff in order to set the key direction(s) for this project.
Again, this was because this project was not about simply improving an existing app, but strategizing the improvements to future-proof bbb Korea’s relevance. We had a long discussion with bbb Korea and narrowed them down to just three to bring them to the next stage.
Testing three idea packages with real users
Now’s the time to test our idea packages: 1. Right information at the right time, 2. Level up, bbb volunteers! and 3. bbb Global. We developed soft-prototyping with images and narratives, and made small booklets to introduce the concepts to the workshop participants — volunteers and potential users.
- Right information at the right time
Right information at the right time is a concept that resolves the many inconveniences found in the current bbb Korea app. Among them are the integration of machine translation, GPS-based location selection, and relevant information package given by the AI after the call. The location information would help volunteers understand the context better and the relevant information package would further help requesters even after the interpretation ends, hence the concept name. bbb interpretation is now at its best with the full advantage of AI and machine learning!
2. Level up, bbb volunteers!
The idea behind this concept is to let people who speak minority languages in Korea like Vietnamese, Thai, and Indonesian — even at a beginner’s level — to be bbb volunteers. On the other hand, major languages like English have a very large number of volunteers, so they can choose to specialize in certain areas, like police or hospital affairs.
Every bbb volunteer starts at the first level, Altruist, and can level up to be an Associate when they have acquired enough experience and are ready to. Volunteers can ask for feedback from requesters after every call about their interpretation and communication skills, and if they wish, they can schedule a mentoring session with an associate.
3. bbb Global
bbb Global is an international platform that provides interpretation between any languages not restricted by time and location. Anyone can be a bbb volunteer and help people from anywhere across the globe, no matter where they are. Volunteers can volunteer at a time convenient for them that may be the middle of the night where the caller is. bbb can completely break down the barriers of location and timezone, making sure people get the help they need.
With these three concepts, we headed to the bbb Korea office in Seoul. We ran a total of four workshop sessions in two days — two in English, one in Russian, and one in Indonesian — each taking up to two hours. For each session, we invited two bbb Korea volunteers and one or two (potential) users of bbb Korea’s services. The aim was to listen to both volunteers and users’ honest opinion about our concepts, discover what was lacking and what we could improve.
In addition to these workshops, we also invited a highly experienced service designer who had extensively worked with services based on natural language processing. All this was more geared towards generating ideas from conversations during the workshop sessions and set key directions for the project, not validating the concepts and ideas we created thus far.
To our surprise, none of the bbb Korea volunteers were in favor of Level up, bbb volunteers! They told us that they are volunteering out of goodwill, so getting reviewed by the people for leveling up seemed… not right. Also, we found out that bbb is in a situation to expand its service for global reach, so bbb Global was also not the scale we should go after.
Still, one very interesting insight came from a bbb Korea volunteer that works in a hospital. The issue was about how interpretation cases in hospitals are often more complicated and thus need an interpreter that is prepared to interpret the situation. The current bbb interpretation system doesn’t really work for them because 1) connection to a certain interpreter is not guaranteed, which can upset both doctors and patients, and 2) some cases can’t be solved with just one call and need to be continued with the same interpreter, which is currently not really available. This insight became the basis for one of our final concepts.
The discussion with experienced service designer was also very useful. She told us that “machine translation is so good already and it’s also more convenient for users. bbb Korea cannot and does not need to compete with it. One day machine translation will replace 99% of community interpreting which bbb Korea is handling at this moment”. This discussion made us think about the new target, new market, new context, new… everything for the future of bbb Korea. This (re)confirmed our goal — helping bbb sustain itself in the ever-changing technological world, and to not compete with technology but to adapt and utilize it to further improve their services. We were on the right track.
Final run
We had two concepts for our final outcome. We called one of them bbb As-is — improving the current bbb Korea service with machine learning and AI. The second one was a completely new service and business concept that we won’t cover in this post.
With bbb As-is, bbb Korea’s service is improved and augmented with the use of A.I. to better meet the needs of both volunteers and requesters. The requesters now have access to not only interpretation requests, but also machine translation through the bbb Korea app. They can now see the time left to be connected to the same volunteering interpreter real-time — set as 30 minutes as default — and extend it if needed with a simple touch of a button. They now also have more chance to be connected to the same volunteer and can express their gratitude by leaving a thank you note to the volunteer.
With keyword input by requesters and the GPS-location information, the app delivers contextual information to bbb Korea’s volunteers before they pick up the phone and provide a more efficient and effective conversation. After the call, volunteers can add keywords that describe what the call was about and re-listen to the audio recording within a set time period — for instance 72 hours if they want to see if there was any mistake or improve their skills. They can leave a message to the users if they would like to help them with the situation, such as the address or phone number they found on the Internet after the call.
Next steps
There are lots of considerations to be made when realizing the ideas we created. For example, the current bbb interpretation service runs on both app call and regular phone call, so shifting to completely app-based might be challenging. There is a long way ahead, but we sure look forward to seeing these concepts realized. But a good news is that our collaborator, EXPC is now working to turn our ideas into reality. Looking forward to it!
Through this intensive project, we learned so much about community interpreting volunteerism, the operation of bbb Korea as a cultural and linguistic NGO, the life of amateur and professional interpreters, as well as hospital staff and police. As always, these learnings often go beyond what we could express in a short presentation or article like this. We are only grateful for these experiences.
We started the project without any idea of how to approach the challenge and what our final outcome would be, but we’ve come a long way and completed it successfully. Huge thanks to bbb Korea and everyone who participated in the interviews and workshops for making this happen.
Written by Kezia Odelia and Seungho Park-Lee
