Sovereign AI
Building Canada's technological independence in artificial intelligence — so our data, our models, and our future stay in Canadian hands.
Canada cannot afford to depend on foreign AI
Nations around the world are racing to build AI systems that reflect their values, protect their data, and serve their citizens. The European Union, the UAE, India, and Japan have all launched sovereign AI initiatives. Canada — despite being the birthplace of modern deep learning — risks falling behind.
Our government data, healthcare systems, legal documents, and public services increasingly depend on models built outside Canada, under foreign laws. When sensitive Canadian information is processed abroad, Canadians lose control over how it is used, stored, and governed. This is not just a technology problem — it is a matter of national sovereignty.
The good news: open-source AI has made sovereignty achievable. Training costs have plummeted, and fine-tuning existing models for Canadian contexts is now both affordable and effective. The window is open — but it will not stay open forever.
Our initiatives
Canadian Data Sovereignty Framework
Developing guidelines and best practices for processing sensitive Canadian data with AI systems that respect PIPEDA, the Official Languages Act, and provincial privacy laws.
Open-Source Model Adaptation
Research into fine-tuning leading open-source models (DeepSeek, Llama, Qwen, Mistral) for Canadian legal, healthcare, and government applications using Canadian datasets.
Compute Infrastructure Assessment
Mapping Canada's current AI compute capacity and identifying gaps that need to be filled for sovereign AI deployment at scale.
Policy Research & Recommendations
Publishing research that informs Canadian policymakers on AI regulation, data governance, and strategic investment in domestic AI capabilities.