Group of Yugambeh Aboriginal warriors dance.
As languages are becoming extinct at an alarming rate, speakers of endangered languages are turning to technology in a race against time to pass on their unique languages and cultures to the next generation.
The United Nations has declared 2019 as the International Year of Indigenous Languages in an effort to promote awareness of the plight of languages that are in danger of disappearing. “Through language, people preserve their community’s history, customs and traditions, memory, unique modes of thinking, meaning and expression. They also use it to construct their future. Language is pivotal in the areas of human rights protection, good governance, peace building, reconciliation, and sustainable development”: all core aspects of the United Nations’ Sustainable Development Goals (SDGs).
Thanks to the benefits of artificial intelligence for language documentation and learning, AI is becoming more important than ever in the fight to save endangered languages.
Jason Lovell was learning the Indigenous language of New Zealand, te reo Māori, but didn’t have anyone to practice with – so he co-founded a Facebook Messenger chatbot powered by IBM Watson artificial intelligence that understands and replies to users in both te reo Māori and English. Even if users make typos or spelling mistakes, “Reobot,” Lovell’s chatbot, has no problem understanding. Lovell hopes to introduce pronunciation help in the near future. By providing learners with the opportunity to converse in te reo Māori no matter where they are during the day, Reobot can help learners develop confidence and skills more quickly.
To teach Indigenous languages of Australia to children living in remote communities, a team of researchers at the ARC Centre of Excellence for the Dynamics of Language (CoEDL) has developed Opie, a low-cost, easily transportable robot. Opie’s eyes engage with the children as they learn through stories, games, and lessons, and the robot records the children’s language skills for teachers to track their progress. CoEDL has partnered with Google to transcribe and build AI models for Indigenous languages, thanks to Google’s open-source AI platform called TensorFlow. This machine learning technology saves linguists literally millions of hours of transcribing CoEDL’s repository of over 40,000 hours of recordings.
Maintaining digital language data and making it accessible for Indigenous communities is a challenge for many organizations. First Peoples’ Cultural Council (FPCC), in its mandate to support the revitalization of Indigenous languages, arts, culture and heritage in British Columbia, works with local communities to archive linguistic data and produce teaching programs and apps through its First Voices platform. FirstVoices’ latest innovation is a Keyboard App that enables users to type in over 100 Indigenous languages on any app in their mobile device, including social media, e-mails and word processing. First Voices stores its data in Nuxeo, an open-source, cloud-native content services engine that incorporates AI and machine learning.
Futurist Thomas Frey envisions a Global Language Archive as a living museum, “The Louvre of Languages,” where even extinct languages can be learned: “Inputs will involve the collection of sufficient video, audio, and written documents for an AI Language Recreation Engine to generate a functional three-dimensional avatar capable of teaching the language to someone wanting to learn it.” The AI Engine would go even further and fill in any linguistic gaps, create a written form of the language if needed, and provide translations between languages. Frey cites the Endangered Languages Project, managed by First Peoples’ Cultural Council and the Endangered Languages Catalogue/Endangered Languages Project (ELCat/ELP) team at University of Hawaiʻi at Mānoa, as the first step towards creating a Global Language Archive. The Endangered Languages Project compiles resources and information on endangered languages thanks to worldwide collaborators, and so far has data on 3418 languages, many of which are represented on their world map of endangered languages.
I founded the AI for Good Global Summit, which provides a platform for dialog on the beneficial uses of AI. The summit was launched in 2017 by the International Telecommunication Union in partnership with sister UN agencies and XPRIZE. The 2018 Summit with the ACM also as partner, focused on practical applications of AI for improving our world, and the 2019 Summit will be held from 28-30 May 2019 in Geneva, Switzerland.
Have a project that applies AI for Good? Consider submitting your work to ITU’s AI Repository to not only gain global recognition and connect with AI stakeholders, but also become eligible for the World Summit on the Information Society (WSIS) Prizes. The 2018 prize-winning projects can be viewed here.