Indigenous Language Revival Is Accelerating—What Is Actually Working?

Canada’s Indigenous-language revival is producing measurable gains in some communities, but the national picture remains fragile. Statistics Canada’s latest detailed census analysis shows that 243,155 people reported being able to conduct a conversation in an Indigenous language in 2021, while the proportion learning an Indigenous language as a second language increased substantially. At the same time, many languages continue to lose speakers, and the number of people reporting an Indigenous mother tongue has declined. The most important lesson is that “revival” is not one national trend: some communities are rebuilding intergenerational transmission while others are still trying to prevent further loss.

The strongest evidence so far points to a combination of approaches rather than a single technology or program: sustained immersion beginning in childhood, community-controlled teacher training, direct relationships between Elders and learners, language use at home and in public life, and digital tools that extend—not replace—human speakers. British Columbia provides one of the clearest examples: its First Peoples’ Cultural Council supported 9,126 people through language programs and funded 3,679 people to participate in immersion activities in 2024–25. Meanwhile, federal researchers have developed speech, grammar, dictionary and text-to-speech technologies now being adopted by more than 25 language communities.

The first surprise: revival is happening even while the national numbers fall

Canada is home to more than 70 Indigenous languages, representing very different levels of vitality. The 2021 Census recorded a 4.3% decline in the number of Indigenous people able to conduct a conversation in an Indigenous language compared with 2016—the first comparable census-period decline since 1991.

That figure can make language revitalization appear unsuccessful. It is not, however, the whole story.

The same census found that the share of Indigenous-language speakers who had learned their language as a second language increased from 25.7% in 2016 to 29.5% in 2021. In other words, a growing proportion of speakers are people who did not acquire the language as their childhood mother tongue.

That matters because revitalization often has to rebuild a broken transmission chain.

For generations, colonial policies—including residential schools—actively suppressed Indigenous languages. A community that once transmitted its language naturally from grandparents to parents to children may now have older fluent speakers, middle generations with varying levels of knowledge, and younger people beginning as second-language learners.

A successful program therefore may initially produce more learners before it produces more first-language speakers.

The numbers differ dramatically by language

The contrast between languages is one of the clearest reasons national averages need caution.

In 2021, speakers of Mi’kmaq increased by 8.0% from 2016, while Blackfoot speakers increased by 19.1%. Michif speakers increased by 44.9%. Some smaller British Columbia languages showed even larger proportional increases: Statistics Canada reported a 173.9% increase for Heiltsuk and a 57.1% increase for Haisla between 2016 and 2021.

Other languages moved in the opposite direction. Dene speakers declined 10.9%, Tłı̨chǫ declined 10.0%, and Slavey-Hare languages declined 20.3% over the same period.

The practical implication is important: there is no single Canadian language-revival strategy that can simply be scaled identically across every community.

A language with thousands of speakers needs different priorities from one with a few dozen fluent Elders.

What appears to be working

1. Immersion is doing something ordinary classes cannot

The most promising approach is also one of the least technologically glamorous: making the language part of everyday life.

A weekly language class can teach vocabulary and grammar. Immersion attempts to make the language the medium through which children learn mathematics, relationships, culture, stories and the ordinary routines of childhood.

That difference matters because conversational fluency requires much more than recognizing words.

British Columbia has invested heavily in community-based language revitalization. The First Peoples’ Cultural Council reported more than $15.1 million in language-revitalization grants during 2024–25, supporting 9,126 people through language programs and 3,679 through immersion activities.

Those figures measure participation rather than long-term fluency, so they should not be interpreted as proof that thousands of people have become fluent. But they provide evidence of expanding infrastructure for learning.

There are also signs that some communities are rebuilding the teacher pipeline required to sustain immersion.

In 2025, Stz’uminus First Nation launched a community-based Indigenous language revitalization program in Hul’q’umi’num’ with the University of Victoria. Its inaugural cohort had 17 learners and the program is designed to ladder from a certificate into a diploma and Bachelor of Education. Mowachaht-Muchalaht First Nation launched a comparable Nuu-chah-nulth program with UVic.

That model addresses a fundamental problem: a language cannot become a sustainable language of education if there are not enough people qualified to teach it.

2. Teacher training is becoming part of language revitalization itself

For endangered languages, teacher education cannot simply mean taking an existing teaching degree and adding a language course.

Communities need educators who understand both pedagogy and the language—and who can operate within community cultural protocols.

British Columbia’s newer programs deliberately bring post-secondary education into First Nations communities. The province reported that three new programs were expected to enroll approximately 50–60 students beginning in 2025 and early 2026, with graduates potentially entering the workforce from 2028.

The province also supports six Indigenous Language Fluency Degree program sites. Government estimates acknowledge that community-based delivery is more expensive than campus-based education and that additional funding is needed to fully support the programs.

This exposes one of the less visible constraints on revival.

A language program may have demand, willing learners and fluent Elders, yet still struggle because there are too few trained teachers, insufficient funding, inadequate learning materials or no pathway from beginner classes to professional-level language education.

3. Elder–learner relationships remain irreplaceable

Technology can record a fluent speaker. It cannot recreate the social environment in which that speaker acquired the language.

That distinction is particularly important because many Indigenous languages have very small populations of fluent first-language speakers. Canada’s National Research Council notes that the majority of the country’s roughly 70 Indigenous languages now have fewer than 500 fluent speakers, many of them elderly.

This is why mentorship and apprenticeship programs remain important.

The strongest models treat Elders and fluent speakers not merely as sources to be recorded but as active participants in teaching.

That changes the economics of revitalization too. Paying knowledge keepers to teach, mentor and develop materials allows expertise to be transferred while reducing the burden placed on a small number of fluent speakers.

It also produces something a dictionary cannot: the ability to understand when and why a particular expression is appropriate.

4. “Silent speakers” are becoming an overlooked resource

One of the most interesting developments in recent research is the recognition of people who understand an Indigenous language but no longer speak it fluently.

Statistics Canada’s 2026 analysis identifies these people as “silent speakers.” In the 2021 Census, approximately 14,020 people had an Indigenous mother tongue but could no longer speak it well enough to conduct a conversation.

This group may have considerably more linguistic knowledge than a conventional speaker count suggests.

Someone who understands a language can potentially verify pronunciation, recognize vocabulary, follow traditional speech patterns and relearn productive speaking skills faster than a complete beginner.

For revitalization programs, that creates a different strategic question:

Should success be measured only by how many fluent speakers exist, or also by how many people can be brought back into active use?

For highly endangered languages, recovering partially retained knowledge may be considerably faster than creating fluent speakers from scratch.

5. Technology works best when it solves a specific bottleneck

The Indigenous-language technology boom is real, but it is easy to misunderstand what AI and language software can accomplish.

The National Research Council’s Indigenous Languages Technology Project has developed verb conjugators, dictionaries, speech-recognition systems, text-to-speech tools and other resources in collaboration with Indigenous language experts and communities.

These tools target practical problems.

Many Indigenous languages are polysynthetic, meaning a single word can contain multiple meaningful components. That makes conventional language-learning and speech-recognition systems designed around English-like word structures less effective. NRC researchers are therefore investigating approaches based on morphemes and other linguistic units.

A digital verb conjugator, for example, can allow a learner to experiment with forms that would otherwise require an instructor’s time.

Text-to-speech can give students additional opportunities to hear pronunciation.

Speech recognition can help researchers search enormous collections of recordings that previously had to be examined manually.

The important point is that technology is reducing bottlenecks around scarce human expertise; it is not replacing that expertise.

Where Indigenous-language technology is already moving

NRC reported that 25 Indigenous language communities had adopted methodologies or technologies developed through its Indigenous Languages Technology Project by 2024–25, exceeding its target of 24 communities by 2027.

That is a meaningful implementation measure—but it is not the same thing as proving that technology has increased community-wide fluency. NRC itself cautions that long-term impacts remain too early to assess for some projects.

6. Digitizing old recordings may be as important as creating new apps

One of the most consequential applications of technology may happen behind the scenes.

Canada has tens of thousands of hours of recordings of Indigenous languages, but much of that material has not been transcribed or indexed. Researchers are developing systems that can segment speech and eventually make recordings searchable.

That could transform language research.

Imagine being able to search an archive for every recorded example of a particular grammatical construction or word rather than asking a researcher to listen manually through hundreds of hours of recordings.

For languages with relatively few remaining fluent speakers, those archives can represent an extraordinary linguistic resource.

But digitization raises an equally important question: who controls the data?

The strongest projects are being built around Indigenous participation and data ownership rather than treating recordings as an unrestricted raw material for AI companies or academic institutions.

That principle becomes more important as AI systems become capable of generating speech, translations and educational material.

7. Inuktut shows the value—and difficulty—of language policy

Inuit Nunangat provides a different model because Inuktut already has substantial numbers of speakers and official-language status in Nunavut.

Yet even here, revitalization cannot be taken for granted.

In Nunavut’s 2025 Language of Instruction Regulations, the government established a clearer framework for bilingual education using Inuktut and either English or French. Community-level District Education Authorities determine which bilingual model applies, reflecting local priorities and teaching capacity.

The models vary substantially. One, called Siku, provides 100% Inuktut instruction from kindergarten through Grade 3 before gradually introducing English. Another maintains 85–90% Inuktut in the early grades, while a third uses a lower proportion.

This is an important lesson for other revitalization efforts: language policy has to be connected to actual staffing and classroom capacity.

A law can declare that a language should be taught. A school still needs teachers who can teach it.

Inuit Tapiriit Kanatami’s current strategy therefore emphasizes bilingual educators, Inuktut learning materials, online resources and a unified writing system alongside broader language policy.

What the federal government is doing—and where the gaps remain

The federal Indigenous Languages Act, passed in 2019, created the Office of the Commissioner of Indigenous Languages and established a framework intended to support Indigenous peoples in reclaiming, revitalizing, maintaining and strengthening their languages.

Government spending has subsequently become substantial. Canada’s 2026 reporting says more than C$1.4 billion will have been invested from 2019–20 through 2028–29, alongside C$162.3 million in ongoing funding.

The government is also funding regional initiatives. In March 2026, for example, Ottawa announced C$4.8 million for five new Northwest Territories projects, including early-childhood and adult immersion, social language activities and multimedia content.

But funding totals should not be mistaken for proof of revival.

Canada’s own reporting acknowledges challenges around adequate funding, capacity and the need for clearer, co-developed measures of progress.

The federal review of the Indigenous Languages Act is another important development. The government began the process in September 2025 and was still advancing it in 2026.

The question is increasingly moving from “Will Canada fund Indigenous languages?” to “Is the funding structured in a way that allows communities to build institutions that survive from one generation to the next?”

What should count as progress?

Speaker counts remain useful, but they are an incomplete measure.

A language can have fewer total speakers while simultaneously becoming stronger in other respects. Conversely, a language can gain learners without yet becoming secure.

A more useful assessment looks at several indicators together:

• Children speaking the language: Are younger speakers beginning to outnumber older speakers?
• Intergenerational transmission: Are parents using the language with their children?
• Home use: Is the language being spoken outside classrooms?
• Immersion: Are learners receiving sustained instruction rather than occasional lessons?
• Teacher capacity: Are new fluent teachers being trained?
• Community use: Is the language used in ceremonies, media, workplaces, government and public life?
• Second-language acquisition: Are learners progressing from introductory courses toward conversational fluency?
• Fluency depth: Are learners becoming capable of extended conversation rather than simply answering standardized survey questions?
• Data sovereignty: Do communities control recordings, dictionaries and digital language resources?
• Institutional continuity: Does a program have stable funding and trained personnel beyond a short grant cycle?

Statistics Canada provides an encouraging example of why multiple indicators matter. Among First Nations people in Atlantic Canada, children aged 14 and younger were more likely to speak an Indigenous language than adults aged 65 and older in 2021. Quebec showed an even larger age reversal among First Nations speakers.

Those patterns are potentially more significant for long-term revitalization than a short-term increase in adult learners because they suggest that transmission may be rebuilding in particular communities.

The urban challenge is becoming impossible to ignore

Language revival cannot be designed exclusively around reserve or remote-community geography.

In 2021, 57,120 Indigenous language speakers—23% of all Indigenous language speakers in Canada—lived in a census metropolitan area or census agglomeration.

That creates a different set of requirements.

An Indigenous person living in Winnipeg, Edmonton, Vancouver or Toronto may be geographically separated from fluent speakers and community institutions but still want to learn the language of their Nation.

Online courses, digital dictionaries, community organizations and urban language circles can therefore provide an important bridge.

But the same research reveals a challenge: self-rated proficiency was substantially lower among Indigenous language speakers living in urban areas than among those outside metropolitan areas.

That suggests that access to a digital lesson is not equivalent to access to a speech community.

The long-term solution may therefore be a hybrid model: technology makes materials and instruction available, while regular gatherings and relationships create the conversational environment needed to turn knowledge into use.

What technology cannot fix

The temptation with AI is to see language revitalization as primarily a data problem.

It is not.

A sufficiently large dataset can help a computer generate a sentence. It does not automatically create a community in which children hear that language from their grandparents, parents, teachers and friends.

There are also risks in building AI systems for languages with small speaker populations. Incorrectly generated forms can spread rapidly if learners assume that a machine-generated sentence is authoritative. Cultural knowledge may be context-dependent or restricted. Recorded voices and linguistic data may have ownership and consent requirements that differ from conventional public datasets.

NRC’s current work explicitly recognizes these issues. Its 2025–26 plan includes research into detecting and watermarking AI-generated Indigenous-language content to help combat fraudulent learning materials.

That is a revealing development. The technology is becoming useful enough that the problem is no longer merely whether machines can generate Indigenous languages, but whether people can reliably distinguish community-validated language resources from synthetic or incorrect ones.

The strongest model is not “old versus new”

The evidence increasingly points toward a layered approach.

An Elder provides pronunciation, cultural context and living knowledge.

A trained teacher turns that knowledge into sustained instruction.

An immersion school gives children repeated exposure.

A family makes the language part of everyday life.

A dictionary captures vocabulary.

A verb conjugator provides practice.

Speech technology supplies additional pronunciation opportunities.

Automatic transcription makes old recordings searchable.

Government funding helps maintain the infrastructure.

Community governance determines what should be recorded, shared and protected.

Each component solves a different problem.

That is why the most promising language-revival programs tend to be ecosystems rather than apps.

What the next decade will reveal

The most important test will be whether today’s growing pool of learners becomes tomorrow’s generation of fluent speakers.

Canada’s current numbers contain both warning and evidence of possibility. Overall speaker numbers declined in the latest census period, but second-language learning expanded, some languages recorded substantial growth, and several communities show encouraging patterns among younger speakers.

The next census will therefore be important, but even it will not answer everything.

The deeper test is whether children who are learning Indigenous languages today continue using them as teenagers and adults—and whether they eventually raise children who acquire those languages naturally.

If that happens, revitalization will have crossed its most important threshold: the language will no longer depend primarily on programs to survive because the community itself will once again be transmitting it.

That is the measure that matters most. Apps, AI systems, grants, dictionaries and government policies can accelerate the journey, but a language is truly being revived when people choose to use it with one another—and when the next generation learns it not simply as a subject, but as part of ordinary life.