It’s fair to assume the government’s AI strategy will try to be a lot of things to a lot of people. The harder question is what it actually commits to. 

To understand what’s different this time, it helps to remember what came before. The Pan-Canadian AI Strategy (2017), was deliberately narrow: focusing on research and talent – steering clear of regulation, deployment and left a glaring gap for commercialization and adoption in the country, 

The attempt to fill that gap came later, through the Artificial Intelligence and Data Act (AIDA, part 3 of Bill 27) – which was effectively killed when government dissolved in early 2025. Canada enters this moment having tried research-first and watched its one regulatory attempt collapse before passing.

That experience has shaped what the Carney government is now trying to do, and what makes it different this time around. Canada’s first Minister of AI and Digital Innovation, Evan Solomon, launched a 30-day public consultation in October 2025, resulting in a public summary report that outlined five key themes: ethical/safety-focused research, transparent governance and risk-based regulation, sovereign AI infrastructure and IP protections, national literacy, and stronger AI security and liability laws. It’s a lot.  Those themes began to take shape in the Spring Economic Update tabled on April 28, gathered under the banner “AI for All.” The six pillars set up the scaffolding ahead of next week’s strategy, supposed to be the building. Whether it’s substance or branding is exactly what’s worth watching.

So, what should you actually be reading for when it drops? A strategy this broad will inevitably say something agreeable about everything – but the signal is in the specifics – the commitments, the timelines, the things left unaddressed. Four areas have dominated stakeholder conversation and each one feels slightly unique:

1. Does the document include a legislative timeline or does it kick the bill further into 2026/2027. Why this matters: A privacy and data update is widely expected, but the signals point to AI -specific governance being handled through amendments to existing frameworks (privacy law, sectoral regulators, etc.) rather than standalone AI act. In June, Solomon said that Bill C-27 was “not gone,” but needed to be “re-examined”, with regulation “taken in steps, starting with privacy and data concerns.” Privacy first, AI second – if at all as a separate piece. Without a binding legislative deadline, the strategy functions as an industrial policy document rather than a governance framework: the former treats AI as an economic opportunity that needs runway, the latter as something to be controlled. A year ago, Solomon was quoted suggesting Canada would move away from “over-indexing on warnings and regulation” to make sure the economy benefits from AI.  The instinct is known, what’s missing is the timeline.
2. Does the Sovereign Compute Strategy get folded in as a pillar or kept separate. Why this matters: The architecture for the strategy is already there from Budget 2024: $2B over 5 years, with its own program and its own funding envelope. It is now being branded as Pillar 4 of the National AI Strategy, with data centre funding grandfathered in. The newer layer worth watching is the Strategic Partnership Framework and Solomon’s non-binding MOUs with industry – the TELUS data centre announcement being the first. Whether the strategy goes beyond infrastructure to address legal sovereignty (CLOUD Act, data residency, model and training data control), and what those MOUs actually require, will tell you whether Pillar 4 is substantive or a rebranding exercise.
3. How it handles the “high-impact” question that broke AIDA — sector-by-sector or horizontal Why this matters: This is likely where it will face the most criticism. The consultation flagged transparent AI governance and risk-based regulation as a key theme – implying movement towards a more strict EU-style tiering. But the strategy is still being released by Solomon as AI/innovation Minister, not by Justice, which structurally pulls it towards horizontal framing over sector-by-sector rules. Expect heavy stakeholder reaction either way.
4. Framing around federal government use of AI in public services. Why it matters: 

Civil society and unions will fight hardest here – and not just over the public service. The National AI Strategy is worker- and economy-facing, which puts it squarely in the middle of a much bigger debate about what AI does to jobs, dignity, and economic security. The conversation shifts are rapid – as recent as Pope Leo XIV’s first encyclical released last week. Against this backdrop, was Solomon’s “pro-worker AI” messaging just inoculation, or does that strategy carry real commitments on retraining, transition supports, guardrails for how employers deploy AI? 

Finally, there’s a fifth that sits outside the document — and probably can’t answer on its own. The path is already being laid: Pillar 6’s focus on global alliances, the Canada–Germany Sovereign Technology Alliance, and the Cohere–Aleph Alpha merger it enabled all point to a middle-power play between the deregulating U.S. and the locked-down EU. The more interesting question is whether it works — whether Canada can carve out a market for those stuck between a hyperscaler and a hegemon, or whether the middle gets squeezed out entirely.

All to say, there’s plenty tow watch. After months of “light, right and tight” – many are eager to see what this actually means for Canadians.

Megan Buttle is the president of data, digital and design of Earnscliffe Strategies.

The views, opinions and positions expressed by all iPolitics columnists and contributors are the author’s alone. They do not inherently or expressly reflect the views, opinions and/or positions of iPolitics.