As artificial intelligence becomes the backbone of daily life, a critical question emerges for Canada: what happens when the systems guiding a nation don't truly understand its people? According to technology investor Harry Glorikian, writing in the Calgary Herald on January 17, 2026, the risk of cultural erosion is real and pressing if Canada does not develop AI that speaks its languages.
The Risk of a Foreign Digital Overlay
The large language models (LLMs) that are increasingly mediating how citizens access information, make decisions, and interact with services are predominantly built by American companies. These models are trained on datasets heavily weighted toward American English and perspectives. Glorikian warns that if these systems lack an understanding of Canadian idioms, customs, and norms, they will quietly rewrite the nation's culture. This issue of "digital sovereignty" is not theoretical dystopian fiction but a tangible threat discussed in a recent Canadian government white paper.
For a country that is officially bilingual and home to more than 70 distinct Indigenous languages, the flattening of this rich tapestry into a generic North American dialect is a significant loss. The consequences for Canadian users will manifest as blind spots in both translation and cultural intelligence. This could lead to awkward educational tutoring from AI, off-key mental health guidance, misfiring customer service chatbots, and public service interfaces that feel alien to local sensibilities.
Canada's Linguistic Diversity Under Threat
The influence of foreign AI models poses a direct threat to Canada's unique linguistic landscape. Dialects from Newfoundland and New Brunswick, along with the bilingual register common among Quebec residents, risk being downranked or treated as "low-quality" by the automated filters of these LLMs. As these models become the default digital teacher, therapist, and business partner for citizens, Canada's national voice could begin to disappear from the machines mediating everyday life, draining the country of its cultural distinctiveness.
Other nations are already taking action to preserve their digital cultural sovereignty. France developed its Mistral model, the UAE created Falcon, and countries like India, Japan, and Nigeria are fine-tuning open models to reflect their national cultures. Canada's discussion, however, has largely centered on computational infrastructure—or "compute." The federal government's new Sovereign AI Compute Strategy, investing billions in data centres and a national AI supercomputer, is a welcome step but does not address the core question of what models will run on that powerful hardware.
The Path Forward for Sovereign Canadian AI
The absence of a robust legal framework compounds the challenge. The proposed Artificial Intelligence and Data Act died on the order paper in January, leaving Canada with only a voluntary generative AI code of conduct and plans for an AI Safety Institute. This regulatory gap highlights the urgency for a coordinated national effort.
The call to action is clear: building computational capacity is only half the battle. Canada must prioritize the development or significant adaptation of AI models that are trained on and reflect Canadian English, Canadian French, and Indigenous languages. This requires a concerted effort to create culturally relevant datasets and to ensure the AI systems of tomorrow are built with a deep understanding of the people they will serve. The future of Canada's cultural identity in the digital age may well depend on it.
