CLEAR Global

TWB’s goal is to break down language barriers that impede people from getting the information they want and need in a language and format that they understand. We focus on changing the communications power dynamic, shifting control to those who speak marginalized languages.

We find solutions at the intersection of language, technology, and the humanitarian and development world. We do this in three ways:

- Building language and communications technology to help people give and get information in a language and format they understand. We build conversational assistants using natural language processing, neural machine translation models, on- and off-line glossaries, and off- to on-line communications and feedback mechanisms.
- Translating, subtitling, transcribing, and providing audio translations of content for our NGO partners. TWB has a community of over 30,000 linguists across the globe who provide these services remotely.
- Working with NGO partners in countries in crisis, such as Nigeria, eastern Democratic Republic of Congo, and Bangladesh, providing direct language support, socio-linguistic research on languages and language issues, and training of field staff on translation and interpretation skills. At the end of 2019, we had trained over 2000 aid workers.

It is our volunteers and language specialists who make these solutions possible, funded by sponsors, donors, and partners.

TWB was formally incorporated as a 501(c)3 in the United States in 2011, and since that time, with the support of a community of over 30,000 linguists, translates over 25 million words per year, as part of humanitarian crisis response, and health and education programming. TWB uniquely brings together people who are linguists, technologists and aid workers to bring cutting edge thinking and technology to bear for people who most need information. Our team includes specialists in natural language processing, computational linguistics, user-centered design, communications, humanitarian response, socio-linguistics, and linguists.

To date the TWB community has translated 50 million words and counting. Conservative estimates suggest that at least 25 million people have accessed the 120 medical articles TWB simplified and translated into 100 languages for the Wiki Med Project. TWB is also working hard to respond to COVID-19. Already TWB has translated over 2 million words of COVID-19 content into over 100 languages, is mapping languages and literacy in vulnerable regions, creating a multilingual glossary of emergency health terms and advice in 23 languages, and combating misinformation by monitoring social media in 12+ languages.

The organization has assembled and deployed teams to work directly on language capacity in the field, most recently in Bangladesh, Democratic Republic of Congo (DRC) and Nigeria. In Nigeria, for example, TWB supported 25 organizations to translate over 1,600 products totaling close to 600,000 words, as well as conducted trainings in humanitarian interpretation, translation, and data collection for over 500 people. In 2017, 23% of Rohingya refugees felt they had information on what assistance they could receive; one year later, thanks in part to TWB, 84% felt they knew what assistance was available to them. TWB investigated language barriers in areas of DRC affected by Ebola to support responders to engage more effectively with people at risk. TWB is also leading the humanitarian community in helping people to understand their right to assistance free from sexual exploitation and abuse. We have simplified, translated and field verified the six principles for the prevention of sexual exploitation and abuse into 100 languages.

TWB has also launched Gamayun, the language equality initiative, an open source platform intended to apply cutting-edge language technologies to improve rapid communications in languages that are currently marginalized. It is a collaborative effort with tech partners, native-speaker communities, NGOs, and content creators and owners, to use text and speech data to automate translation in under-served languages. Using a repeatable, scalable model, TWB and its coalition of partners are piloting three languages to test and refine the process, with the aim of bringing 10 underserved languages online over the next 5 years.