Canada Has a New AI Council. What Can It Actually Do?
Matthew Leo · Published October 4, 2026 · Canada Tech
Canada now has a National Council on Artificial Intelligence. Prime Minister Mark Carney announced the 13-member group on October 2, saying it will provide independent advice as Ottawa expands its AI for All strategy.
The council may have access to senior decision-makers, but it is not a regulator and it cannot make policy on its own. That distinction matters because the government's announcement gives it a large mandate without giving it formal powers.
What the council is supposed to work on
According to the Prime Minister's Office, the council will advise on AI adoption, Canadian companies, sovereign infrastructure, public services, safety and the protection of democracy. It is also expected to connect people in government, research, business and investment.
The members are Ajay Agrawal, Yoshua Bengio, Sanja Fidler, John Graham, Gillian Hadfield, Jordan Jacobs, Eric Lamarre, Valerie Pisano, Senator Thomas Pitfield, Jody Thomas, Louis Vachon, Michelle Zatlyn and ex officio member Patrick Pichette. The list brings together researchers, investors, technology executives and people with government or policy experience.
The published membership does not reserve seats for labour organizations, consumer advocates, Indigenous governments or civil-society privacy groups. The council can still seek those views, but Ottawa has not explained how outside participation will work.
Advice is not authority
A small Privy Council Office team will support the group. Ministers and departments will remain responsible for decisions and implementation. The announcement does not give the council rule-making, enforcement, spending or procurement authority.
Ottawa also did not publish a meeting schedule, deadline for recommendations, commitment to release minutes, or a requirement to explain when the government rejects the council's advice. Those details will determine whether the public can judge its influence.
This is not necessarily a powerless body. A council with direct access to the Prime Minister can shape priorities before a bill, budget measure or procurement reaches the public. That makes transparency more important, especially when several members work in industries that may benefit from government AI spending.
The targets need to be measured
The AI for All strategy targets an additional $200 billion in economic growth, 250,000 new AI-related jobs over five years and an increase in business AI adoption from just over 12 per cent to 60 per cent by 2034. These are government targets, not independent forecasts.
The council should press Ottawa to define how it will measure an AI-related job, business adoption and the claimed economic contribution. Without a public baseline and regular reporting, a large target can be difficult to test.
Sovereign infrastructure will need the same clarity. Mapletechie recently examined what a proposed Canadian sovereign-AI service does and does not guarantee. Keeping equipment in Canada is only one part of sovereignty. Buyers also need to know who controls keys, administrators, logs, software updates and foreign dependencies.
What to watch next
The next useful information is not another list of ambitions. Ottawa should say when the council will meet, whether recommendations will be public, how conflicts will be disclosed and how people outside the technology industry can contribute.
Until then, the council should be described accurately: it is an advisory group close to government, not Canada's AI regulator and not an independent oversight body.
Sources
- Prime Minister of Canada: council announcement and membership
- BetaKit: independent coverage of the launch
Tags: National Council on Artificial Intelligence, AI for All, Canadian AI policy, Mark Carney, sovereign AI