Government of Nunavut preserving endangered Inuit languages and culture with the help of artificial intelligence and Microsoft

By Kevin Peesker, President Microsoft Canada 

In January 2021, in close collaboration with the Government of Nunavut, Microsoft announced the launch of Inuktitut in Microsoft Translator. Today, at the start of Uqausirmut Quviasuutiqarniq, an annual celebration of Inuktut language in Nunavut, we’re pleased to announce the addition of Inuinnaqtun text translation, along with some key updates to Inuktitut, including updated language models and text translation of Roman orthography to Microsoft Translator.

Inuktitut is the primary dialect of the Inuktut language; it is spoken by approximately 40,000 Inuit across Inuit Nunangat, the Inuit homeland in Canada, and used by 70 percent of Nunavut’s residents. Inuinnaqtun, also a dialect of Inuktut, is on UNESCO’s list of endangered languages with fewer than 600 people who speak it as their mother tongue. It is spoken mostly in the Kugluktuk and Cambridge Bay communities in the Kitikmeot region of Nunavut.

The Government of Nunavut engaged Microsoft on the Inuktitut project as part of their ongoing effort to ensure the vitality of the Inuktut language. Today’s announcement continues to build upon this work, helping to preserve the language for future generations. Since the original launch, feedback has been positive. Community members asked for more updates to make Inuktitut even more accessible and support to help ensure the longevity of Inuinnaqtun.

Tools such as Microsoft Translator help to increase the number of people who are learning and using Inuinnaqtun in daily life, while improving accessibility for residents of Nunavut. For example, unilingual speakers who often travel long distances for healthcare, can communicate more effectively with doctors and other health care providers; younger generations can learn to communicate with elders where they may not know and understand the same language; workers using Microsoft Office can email in their own language and translate emails as they choose.

The addition of Inuinnaqtun to Microsoft Translator was only possible because of the efforts of the community and groups like the Kitikmeot Heritage Society. The Government of Nunavut turned to the community to help build and test the language models needed to make Inuinnaqtun available in Translator. Like all AI-powered tools, it will improve further with use, just as Inuktitut has in the last year, enabling us to deploy important updates.

Preserving Indigenous languages is a critically important step on the road to reconciliation – in fact, the Calls to Action of the Truth and Reconciliation Commission of Canada assert that “Indigenous languages are a fundamental and valued element of Canadian culture and society, and there is an urgency to preserve them.”

This project is the latest in Microsoft’s longtime collaboration with the Government of Nunavut to transform how they deliver services with technology. Nunavut has deployed Microsoft Teams to many of its 5,200-plus employees and additional support workers, providing new ways to connect virtually across its vast territory. This allows government employees and residents to access skills training, and other essential services, potentially saving Nunavut millions of dollars in travel costs. Nunavut also piloted Windows 365 to help securely manage an elastic workforce, enabling them to deliver the day-to-day expertise needed to run the government and improve the quality of life for local communities.

It is an immense privilege to support the Government of Nunavut, departments, staff, and residents in this important work to modernize government infrastructure and to help make work, school and life more accessible for all.

Available in the Microsoft Translator apps, Office, and Translator for Bing, users can use AI-powered Azure Cognitive Services Translator and Azure Cognitive Services Speech to add both Inuktitut and Inuinnaqtun translation to apps, websites, workflows and tools. Text translation is based on neural machine translation models.


Related Posts