MONTREAL — He’s a tech guy, but Ian Rae is also something of a plumber. Sixteen years ago, the Montrealer founded cloud.ca, among the first cloud platforms to be owned, operated and to have its servers physically located in Canada. The idea behind it was instructive: digital sovereignty isn’t just controlling the guts behind tech, but the saleable data sets these guts produce as well. “Sovereignty isn’t only about soil, it’s about supply chain,” Rae told me.

Rae’s soup-to-nuts take applies remarkably well to data-hoovering artificial intelligence. When he delivered it at a roundtable at Montreal AI research institute MILA in September, it got him noticed by the likes of Prime Minister Mark Carney and AI Minister Evan Solomon, who were in attendance.

His reward was a spot on the government’s 26-member AI strategy task force, charged with recommending how best to fund, develop, implement, commercialize and otherwise render safe what could be the most consequential and far-reaching technological development of the last century.

The 26 task force members—divided into eight themes, including “security” and “commercialization of AI,” according to a seven-page briefing I got my mitts on—have 30 days to produce a five- to seven-page memorandum reflecting their wisdom and that of their networks. The government wants to deliver its AI strategy by the end of the year. “I’m not happy about it,” Rae said, of the timeline.

He isn’t the only one. Rae was among a handful of AI stakeholders I spoke with who, while chuffed at being included, questioned the wisdom of limiting discussion on the country’s AI policy to what that government briefing refers to as a “30-day sprint.” They point out how the AI task force is a Tower-of-Babel cacophony of often-conflicting sectors, and will deliver a mess of contradictory recommendations as a result. Another member I spoke with worries the task force itself is but window dressing, and the government’s AI plan is effectively pre-baked.

You certainly can’t accuse Carney and this government of inconsistency in its boiler-room approach to AI legislation. The Liberals have emphasized velocity when it comes to building housing, laying high-speed rail, dissolving trade barriers and killing unpopular or inconvenient taxes. There are a few very good reasons for this, not the least of which is the Gong Show that is our neighbour and largest trading partner. 

Yet there are inevitable speed limits. The feds want to build 500,000 housing units a year, but in Jasper, Alta., they’ve struggled to build fewer than 350 since July 2024, as my colleague Catherine McIntyre recently reported. High-speed rail, built quickly, is a great idea; a scarcity of made-in-Canada rail steel is a reality. The government says Indigenous consent for big infrastructure builds is crucial—but what happens if or when the government doesn’t get it?

In a way, moving quickly on AI makes sense. The lag between mass adoption of the technology—ChatGPT, now three years old, has 700 million weekly users—and government regulations is especially yawning. Social media platforms, having operated without meaningful regulatory oversight for much of their existence, are toxic cautionary tales as a result.

Yet it’s difficult to understate the consequences of getting AI policy wrong. It’s not just the myriad practical and ethical issues AI presents, or that its Big Tech purveyors have largely ignored warnings as to AI’s potential dangers. For Canada, a middle power that punches well above its weight when it comes to AI, it’s also this: the world wants what we have, and it’s vitally important that we protect it. “You need to have unique, high-quality data sets with moats around them, and that you have some particular advantage in generating, managing and then translating into economic advantage,” Rae told me. 

If moats are the metric, the results are so far mixed. The more regulation-minded types on the task force are perplexed as to why a government allegedly obsessed with keeping things Canadian would fork over $240 million for Toronto-based Cohere to buy AI compute space from CoreWeave, a New Jersey-based cloud computing firm. “A dumb, dumb deal,” is how one member referred to the Cohere-CoreWeave deal. 

Of course, this person is on the task force, which suggests the government isn’t allergic to criticism—and isn’t stuck in the same AI echo chamber as, say, the U.S. president.

Judging by what it is asking the task force to do, the federal government seems genuinely concerned about AI’s potential risks, not just its benefits. “What conditions are needed to ensure Canadian AI research remains globally competitive and ethically grounded?” reads a typical question put to task force members.

It’s about as Canadian a question as you’ll find on the topic. If only there were more time to answer. 

Martin Patriquin is The Logic’s Quebec correspondent. He joined in 2019 after 10 years as Quebec bureau chief for Maclean’s. A National Magazine Award and SABEW winner, he has written for The New York Times, The Guardian, The Walrus, Vice, BuzzFeed and The Globe and Mail, among others. He is also a panellist on CBC’s “Power & Politics.”