Skip to content

AI sovereignty in Canada: key takeaways from the Vatican AI Symposium

Levio’s Richard St-Pierre shares a Canadian perspective from the Vatican on sovereign AI, public trust, governance and responsible public-sector adoption.

June 19, 2026

Richard St-Pierre, Quantum and AI Sovereignty Senior Advisor, Levio

When I took part in the international symposium “AI and the Future of Human Dignity” at the Vatican in May, the conversation was not only about technology. It was about society. 

The event brought together 22 select global experts consisting of Nobel laurates, leaders from scientific, economic and institutional circles to examine how artificial intelligence is reshaping work, governance, social sustainability and human dignity. It was a rare setting because it created space for a question that is becoming increasingly urgent: how do we govern a technology that is beginning to act inside our economies, organizations and institutions? 

For Canada, this question is immediate. AI is no longer only a tool used to improve productivity or automate tasks. It is becoming an actor inside institutional life, capable of influencing decisions, executing actions, routing information and reshaping how services are delivered. 

That changes the nature of the responsibility. 

From AI adoption to AI agency 

Many institutions and organizations have approached AI as a technology to adopt. That framing made sense when AI was primarily used to analyze information, support employees or automate repetitive processes. 

Agentic AI changed that conversation. 

Agentic systems act with greater autonomy. They recommend, decide, transact, generate outputs and influence workflows in ways that affect real people and real institutions. In the public sector, that raises deeper questions of accountability, transparency and control. 

Institutions need to ask themselves: 

  • If AI supports a public service, who governs the system?

  • If AI influences a decision, who can explain it?

  • If AI relies on external infrastructure, who controls the environment it depends on?

  • If AI becomes part of a critical process, how does the institution remain accountable to the citizens it serves? 

These are technology questions, but they are also questions of public trust. 

Sovereignty is becoming a question of control 

In Canada, digital sovereignty is often discussed in terms of ownership, geography or data residency. Those factors are part of the equation. But they do not tell the whole story. 

Sovereignty is about meaningful control. 

It is about control over data, infrastructure, security, governance, operating models and decision-making processes. It is about whether a public institution can continue to protect, govern and operate the systems Canadians depend on when the environment changes. 

A system may appear sovereign on paper while still depending on technology, platforms or operating models that limit practical control. For governments across all levels, that distinction matters. 

Sovereign AI requires more than policy intent. It requires architecture, modern data foundations, secure infrastructure, resilient operations, clear governance, cyber readiness and quantum-safe security considerations. It also requires the ability to move beyond pilots and bring AI into real operating environments responsibly. 

Canada does not need to choose between ambition and responsibility. It needs both. 

Canada’s window for responsible adoption 

One of the most important themes from the Vatican symposium was that AI’s impact is no longer theoretical. It is already reshaping economic life, labour markets, social relations and institutional trust. 

For Canada, the challenge is pace. 

AI capabilities are evolving quickly. Public procurement, governance and policy cycles often move slower. That gap creates pressure for public institutions that need to modernize services, improve productivity and build digital capacity while maintaining high standards of transparency, fairness, security and accountability. 

Responsible adoption cannot mean standing still. It means building the capacity to move with discipline. 

That is especially important for public institutions. The public sector carries responsibilities that go beyond efficiency. It must protect rights, preserve trust, deliver services reliably and remain accountable for the systems it uses. 

AI can help governments improve service delivery, reduce administrative burden, support employees and strengthen decision-making. Those benefits are real, but they become durable only when adoption is supported by governance, infrastructure and security models that citizens can trust. 

A third path for middle powers  

At the symposium, I spoke about the role of countries like Canada in a global AI environment shaped by major powers and large technology ecosystems. 

Canada cannot simply copy the dominant model. The level of capital investment, infrastructure concentration and platform power held by the largest global players is difficult for most countries to match directly. 

But that does not mean Canada should accept dependence as the only option. 

There is a third path for middle powers such as Canada: collaboration on sovereign digital infrastructure, trusted data-sharing models and governance frameworks that preserve national control while supporting innovation. 

This is where sovereignty and collaboration meet. 

Countries like Canada can work with trusted partners and like-minded nations to create practical models for sovereign AI. That can include shared infrastructure, secure cross-border data collaboration, interoperable governance frameworks and approaches that allow institutions to innovate while maintaining control over critical systems. 

For the public sector, this is especially relevant. Governments increasingly need to collaborate across borders on defence, space, healthcare, climate, emergency response and economic resilience. The ability to share intelligence, data and digital capacity securely may become one of the defining requirements of public-sector modernization. 

Sovereignty does not have to mean isolation. Done properly, collaboration can become one of the ways sovereignty is protected. 

Why implementation matters 

The Vatican symposium was an important global forum for discussing AI’s human and societal impact. But the next step is execution. 

This is where I believe Canada has an opportunity. 

Canada has strong AI research capacity, sophisticated public institutions, world-class talent and a public-sector tradition grounded in trust and accountability. The next challenge is turning those strengths into operational capacity. 

At Levio, our perspective is practical. As a Canadian-owned and truly sovereign digital transformation and systems integration firm, we see sovereign AI as an implementation question as much as a policy question. The next phase of adoption will depend on whether institutions can move from strategy to architecture, from principles to operating models and from pilots to secure, governed and scalable deployment. For Canada’s public sector, this means building the capacity to govern AI in real environments, protect data and infrastructure, prepare for quantum-era security risks and maintain accountability as these systems become more deeply embedded in the services Canadians depend on. 

Public trust as the measure of sovereign AI 

The future of AI will not be measured only by technological power. It will be measured by whether these systems strengthen human dignity, improve institutional trust and support the public good. 

For Canada, AI will continue to reshape how institutions operate and how services are delivered. The question is whether we can adopt it in ways that preserve control, protect citizens and strengthen confidence in the systems that serve them. 

Sovereign AI is becoming a public trust issue because the systems being built today will shape how decisions are made tomorrow. 

Canada has an opportunity to approach this moment with both ambition and responsibility. That means building the digital foundations, partnerships and governance models needed to adopt AI at scale, while keeping human judgement, public accountability and national sovereignty at the centre.