Access Without Architecture

AI for All, the federal government’s national AI strategy, landed on June 4, 2026 [1]. It promises AI literacy training for a million post-secondary students, and access to a ‘trusted AI agent’ for every student, regardless of discipline [2]. It sounds like the kind of announcement that changes things. Although it made headlines, on closer scrutiny it feels like a gift with no assembly instructions.  

The promise from Ottawa comes with no curriculum standards attached, no dedicated funding,  and no governance architecture to make them real. Education in Canada is constitutionally a  provincial matter. What the federal government has produced, in effect, is a mandate with no  mechanism to deliver it — and universities and colleges are now expected to build that mechanism themselves, on their own time and their own budgets. 

The Promise 

The pitch is genuine enough to sound serious. Beyond the one-million-student target, more than 3,000 educators are to receive AI learning kits. The training will not be limited to, “here’s how to use ChatGPT.” It’s meant to cover bias, misinformation, privacy risk, and unsafe use — the goal being informed participants, rather than passive users of AI [1]. 

I like that the framing is more ambitious than most comparable national strategies but how are we going to meet this expectation? Ambition and infrastructure are not the same things. The plan relies on Canada’s three national AI institutes — Amii, Mila, and Vector — as delivery vehicles [1]. Amii alone currently trains 125,000 students a year. Scaling that to a million is not a matter of enthusiasm; it is a matter of capacity that does not yet exist. 

Canada ranked 44th out of 47 countries on AI training and literacy in a recent global trust study, with fewer than a quarter of Canadians reporting any AI training at all [1]. These numbers underline the stakes clearly, and something needs to be done, but is this the right way forward? 

Ottawa’s limitations  

Here’s what no one is saying aloud – the federal government can’t tell a university to change its curriculum, it can only announce a goal and hope it is acted upon [3]. Universities are left to fund and fill the gap between federal ambition and provincial jurisdiction. There is no joint plan and one policy analyst pointed this out just before the strategy even dropped: Canada needed a national framework and widespread literacy, not one standing in for the other [4]. 

Universities are in a double blind situation where if they move early on AI literacy they might have to pivot later when a national standard is available. A college that waits for provincial direction may simply wait for it to arrive, with no current timeline. Students, meanwhile, arrive on campus already expecting the access the Prime Minister announced [5]. 

The ‘trusted’ agent 

The plan to give every student access to a trusted AI agent sounds foolproof, until you ask the  question… trusted by whom? Who will certify those providers, what does trusted entail and what happens to student data flowing through these systems… who sets the governance standards? 

Most Canadian universities scrambled to have AI use policies, built in years following generative AI’s arrival on campus in 2022 and 2023, largely to protect academic integrity. A review of sixteen of those universities found real inconsistencies between their AI rules and their academic integrity rules [7]. Dropping a federally endorsed access program on top of that patchwork doesn’t simplify anything; it just adds another layer that may or may not agree with what’s already there. 

There is a much deeper issue underlying this, and that relates to sovereignty. The access will run through platforms governed by other countries’ laws, not Canada’s [1]. The strategy’s answer — sovereign compute infrastructure — is five years away. Students get access now.  

Literacy = fluency 

The Strategy should be credited for its literacy language that asks students to recognize bias and misinformation, not just prompt a chatbot well [1]. That’s a real step up. 

However, recognizing that an AI output might be biased is not the same as understanding why, or how, or who is accountable for it. Research on Canadian institutions already shows AI systems shaping decisions on retention, resource allocation, and program planning — decisions that carry real risk of bias and opacity, especially for marginalized students [8]. Spotting a bad output is one skill; questioning the system behind it is another. Right now, the strategy only builds the first. Only 14% of recent grads are confident in applying AI tools in their professional work [9]. 

Holding it together 

Universities and colleges did not design AI for All. They were not meaningfully consulted before it was announced. Yet they are now the default delivery systems for the biggest AI literacy initiative in Canada’s history. There is no dedicated funding for curriculum redesign or faculty training attached [10]. Elsewhere, people have proposed the obvious fixes — oversight committees, campus AI audits, dedicated governance staff [11]. None of that comes funded here. 

The federal instrument that might have set binding national standards, the Artificial Intelligence and Data Act, died in committee in 2025 [12]. What remains — the Treasury Board’s automated  decision-making directive, PIPEDA, a handful of voluntary codes — were never built for a  classroom. Student groups have noticed the tension too, asking for access that expands  opportunity without leaving students exposed [13]. It’s a reasonable ask but not an easy one when access shows up years before the governance arrives.  

Next steps 

None of this makes AI for All a bad idea. It’s the biggest federal statement on AI and higher education Canada has ever made, and it gets literacy right — treating it as more than tool proficiency. The intent is clear but the infrastructure needed to back it up is missing. 

Institutions now need to define ‘trusted’ on their own terms before Ottawa does it for them. They have to line up AI access with the integrity policies they already have, and train faculty not just to use these tools but to explain how they think, where they fail, and who’s supposed to be watching them. 

A national strategy can set direction. It cannot substitute for the slower, harder work of institutions deciding, discipline by discipline, what responsible AI use looks like. That work is ours to do, as Canadians, as institutions of higher education, together— without help.  

This editorial draws on published research and policy analysis; all interpretations and conclusions are the author’s own.

References 

1. Innovation, Science and Economic Development Canada. (2026, June 4). Canada’s national  artificial intelligence strategy: AI for All. Government of Canada. https://ised isde.canada.ca/site/ised/en/canadas-national-artificial-intelligence-strategy-ai-all
2. Prime Minister of Canada. (2026, June 4). Prime Minister Carney launches AI for All:  Canada’s new national artificial intelligence strategy [News release].  https://www.pm.gc.ca/en/news/news-releases/2026/06/04/prime-minister-carney launches-ai-all-canadas-new-national-artificial 

3. Ubyssey. (2026, June 11). Here’s what Canada’s AI strategy could mean for students. https://ubyssey.ca/news/canadas-federal-ai-strategy-for-students/
4. Ben Karroum, W. (2025, September 19). Canada needs a national strategy for generative AI  in higher education. Policy Options. https://policyoptions.irpp.org/2025/09/ai-higher-ed/
5. CBC News. (2026, June 1). Draft federal AI strategy aims to scale up adoption, offer literacy  training by 2031. https://www.cbc.ca/news/politics/federal-artificial-intelligence-strategy 9.7216576
6. CBC News. (2026, June 1). Draft federal AI strategy aims to scale up adoption, offer literacy  training by 2031. https://www.cbc.ca/news/politics/federal-artificial-intelligence-strategy 9.7216576
7. Discourse and Writing/Rédactologie. (2024). Examining AI guidelines in Canadian  universities: Implications on academic integrity in academic writing. https://journals.sfu.ca/dwr/index.php/dwr/article/view/1051 

8. McConvey, K., Ghai, M., Lee, R., & Guha, S. (2026). Risk, retention, and the algorithmic  institution: Artificial intelligence as a policy response to higher education in crisis.  Canadian Public Policy, 52(S1), 155–174. https://doi.org/10.3138/cpp.2025-030
9. Theconversation.ca. (2026, June 11). Canada’s AI strategy and universities: Beyond literacy,  students need to be fluent with AI. https://theconversation.com/canadas-ai-strategy-and universities-beyond-literacy-students-need-to-be-fluent-with-ai-283450
10. Azevedo, R. (2025). Institutional policies on artificial intelligence in higher education:  Frameworks and best practices for faculty. New Directions for Adult and Continuing  Education. https://doi.org/10.1002/ace.70013
11. EDUCAUSE. (2024). AI policy development: An action plan for higher education governance. EDUCAUSE Review.
12. AI Magazine Canada. (2026, April 22). The AIDA Act, explained. https://aimagazine.ca/aida act/
13. Canadian Alliance of Student Associations [CASA]. (2026). Post-secondary students  respond to the federal government’s ‘AI for All’ strategy. https://www.casa acae.com/ai_for_all