Canada's newly appointed UN Ambassador David Lametti says there is broad agreement among countries — including emerging economies in Asia — that artificial intelligence needs safety guardrails. The remarks signal where Canada intends to focus its diplomatic energy as it works to shape how the world governs AI through international bodies.

Lametti took up the ambassadorial post in New York on November 17, 2025, replacing longtime incumbent Bob Rae. Yahoo Finance Canada reported his comments on June 29, 2026. He highlighted the convergence on safety as a meaningful starting point from which more detailed governance rules could be built.

Here is what needs careful unpacking: agreement that AI safety matters does not translate into agreement on what "safe" actually looks like in practice. Countries may share the concern but disagree sharply on who should oversee AI development, how companies should be held accountable if things go wrong, or how to address harms that are harder to pin down — algorithmic bias, for instance, or violations of countries' data sovereignty. In UN negotiations, countries often agree on general principles while remaining far apart on the specific rules that would enforce them. Lametti's framing of safety as a foundation, rather than claiming progress on binding international rules, reflects a realistic assessment of where the UN process currently stands.

Canada has invested in this file for several years. The country has sought to position itself as a credible voice on AI governance since the 2018 Montreal Declaration on Responsible AI. Canada's research strengths in Toronto, Montreal and Edmonton give it the technical credibility to engage in detailed policy debates. Lametti, a former federal Justice Minister and Attorney General with a legal background, brings a regulatory and rights-based approach to the ambassador role rather than a primarily commercial one. That stance tends to resonate in multilateral forums where smaller countries and nations in the Global South worry that governance rules will simply protect the large technology companies that already dominate the market.

The Asian angle matters geopolitically. India, South Korea, Japan, and China — all major AI players — each hold different regulatory philosophies and strategic interests. Japan has favoured voluntary industry guidelines; South Korea has enacted binding rules for specific sectors; India is trying to balance its own desire to set regulations with its ambition to become a major AI exporter. China's positions at the UN tend to emphasise state sovereignty and express skepticism toward transparency requirements, which it sees as Western-led. That Lametti identifies a safety consensus across this diverse group is a diplomatic observation — one about countries finding common ground — rather than a statement that they have aligned on technical or legal questions.

Canada's institutional position shapes how it approaches these negotiations. As a G7 member without a permanent seat on the Security Council, Canada has historically worked through coalitions of medium-sized powers and norm-setting bodies — the G20, the OECD, and UN working groups — to influence global governance. Lametti's predecessor Bob Rae was known for a rights-focused approach to multilateralism. Lametti's background as a former justice minister suggests continuity with that tradition, and it is relevant because AI governance increasingly touches on rule-of-law questions: whether automated systems afford due process, how privacy is protected, and how governments' own use of AI in law enforcement or benefits administration is kept accountable.

The UN is not the only venue where this is happening. The OECD's AI Policy Observatory, the Global Partnership on AI — which Canada helped establish — and the European Union's AI Act all represent parallel diplomatic and regulatory tracks where countries are debating these issues. Canada's resources are limited, and Lametti's remarks suggest Ottawa has chosen to prioritise the UN as the place where the widest coalition, including countries in the Global South, can be included in the conversation. Whether that reflects a strategic judgment about where lasting global standards are most likely to take root remains unclear from what the government has publicly said.

What bears noting: Lametti is presenting safety not as something Canada is trying to export, but as a concern all countries share. In the diplomatic work of building global norms, that approach carries weight. Countries that feel their worries have been genuinely heard — rather than having a finished agreement imposed on them — are more willing to stay engaged through the harder negotiations that lie ahead: decisions about how to enforce rules, who bears legal responsibility when AI causes harm, and what access all countries should have to AI technology.