A Canadian nonprofit organization launched in 2022 has successfully influenced government policy on artificial intelligence governance, demonstrating that smaller nations can meaningfully shape how the world manages existential risks from advanced AI systems. AI Governance and Safety Canada (AIGS), operating largely on volunteer effort and personal debt, has secured seven expert testimonies before Parliament, shaped recommendations adopted by the government, and connected over 1,500 Canadians around AI safety concerns.

How Did a Volunteer-Run Organization Influence National AI Policy?

AIGS Canada's journey began modestly in summer 2022 when founder Wyatt Tessari L'Allié received a small grant to connect scattered AI safety advocates across the country. The organization faced an immediate funding crisis. Traditional AI safety grants from American foundations posed credibility risks in government circles, and charitable status prevented direct political advocacy. After initial fundraising efforts failed to materialize, the organization nearly disbanded.

Instead of shutting down, Tessari L'Allié took on personal debt and refocused the organization's mission. The breakthrough came through policy white papers titled "A Plan for Canada," which translated complex concerns about artificial superintelligence (ASI) into actionable recommendations for Canadian lawmakers. The government adopted the top recommendation from both the 2023 and 2024 white papers, a rare achievement for civil society organizations.

The organization's impact expanded through strategic engagement with Parliament. AIGS submitted more comprehensive recommendations on Canada's AI and Data Act than any other organization, carefully translating technical risks into practical legislative language. One parliamentary testimony went viral on social media, reaching 2.3 million views and 119,000 likes on Instagram.

What Are the Key Mysteries About AI That Remain Unsolved?

While AIGS Canada focuses on governance and policy, fundamental scientific questions about how advanced AI systems actually work remain largely mysterious even to their creators. According to AI researcher Lance Eliot, there are at least ten major unresolved questions about modern AI that experts cannot definitively answer, despite billions of dollars invested in AI development and billions of people using these systems daily.

These mysteries span the core mechanics of how AI learns and behaves. The persistent unknowns create a tension in how society should approach AI risk. Some argue that without complete understanding and control of AI systems, humanity cannot be truly safe. Others contend that society should be satisfied that AI works effectively, even if the internal mechanisms remain opaque.

• The Scaling Mystery: Why does adding more computing power and memory to AI systems produce increasingly intelligent behavior, and is there a limit to this approach, or does it continue indefinitely?
• The Explanation Mystery: Can researchers produce clear, human-understandable explanations of the real-time calculations happening inside AI systems, or will they remain fundamentally opaque?
• The Reasoning Mystery: Are current AI systems actually reasoning in the way humans do, or is it misleading to describe their processes as reasoning at all?
• The Hallucination Mystery: What causes AI systems to generate false or nonsensical information, and can these errors be completely eliminated?
• The Generalization Mystery: Can AI systems apply knowledge beyond their training data, or are they fundamentally limited to patterns learned during development?
• The Goal Mystery: Will AI systems eventually develop their own goals independent of human instruction, or will they always be shaped by human-defined objectives?
• The Existential Risk Mystery: Is humanity ultimately vulnerable to catastrophic harm from AI, or can societies successfully navigate and overcome the risks associated with advanced AI systems?

Eliot noted that many researchers have proposed solutions to these mysteries, but widespread agreement on any definitive answer has not emerged across the scientific community. The persistence of these unknowns matters because they directly affect how policymakers like those in Canada should approach AI governance.

Why Does Canada's Approach Matter for Global AI Safety?

AIGS Canada's work addresses a critical gap in how nations prepare for advanced AI risks. The organization operates from the premise that in a future where AI development goes well, multiple countries and institutions worldwide will have contributed meaningfully to managing risks. Canada's specific contributions could include seed funding for critical AI safety research, hosting national conversations about superintelligence that shape global narratives, or leveraging diplomatic leadership to coordinate international AI governance.

The organization's success despite severe resource constraints suggests that effective AI safety work does not require massive budgets or institutional prestige. AIGS Canada operated with just one unpaid staff member and over 700 volunteer sign-ups, yet influenced parliamentary testimony, media coverage in major outlets including CBC The National and Radio-Canada, and government policy adoption.

This model mirrors work being done by organizations in other jurisdictions, such as the Future Society in the European Union and the Centre for Long Term Resilience in the United Kingdom. However, AIGS Canada remains the leading civil society organization doing this work in Canada, and among the first to establish such efforts in any country.

The broader context underscores why AIGS Canada's work matters. Despite growing numbers of Canadian parliamentarians and government officials being briefed on superintelligence risks and expressing genuine concern, AI safety has not yet become a political priority in Ottawa. AIGS Canada's strategy of translating abstract existential risks into concrete policy recommendations appears designed to bridge that gap.

As AI systems become more powerful and their impacts more visible, the organization's potential to influence Canadian policy and contribute to global AI governance frameworks will likely grow. The organization remains focused on the question that launched it four years ago: "What can we do in Canada, and from Canada, to ensure positive AI outcomes?"