Is Canada Ready for the Great AI Disruption?

Nope.

No one knows for sure whether major society-wide changes driven by AI-related technologies are imminent or decades away…or nothing but empty hype. We’re similarly in the dark over the scope of whatever change does come. But it sure would be nice to have some kind of national policy response in the works just in case.

Sure, the Liberals are hyper-focused on making sure no Canadian will ever be able to use AI to generate something that doesn’t meet CRTC approval. And they’re busy enough planning their lottery ticket purchases major investments into AI research and infrastructure build-out. But I haven’t yet seen evidence that the government has been serious about gaming worst-case scenarios and thinking about how the disruption could be managed.

So this post will offer a couple of realistic scenarios detailing what could happen, followed by some practical suggestions for policies that could possibly lead us to a softer landing.

The “Gee, AI Sure Is Helpful” Scenario

On the plus side, we could experience a relatively modest expansion of AI integration that - through cost reductions and new digital services - boosts real gross domestic production (GDP) growth to 2.5 percent annually (compared with a current baseline of around 1.8 percent).

The flip side to that is the downscaling of the current workforce by 10 to 20 percent. We’d see some wage polarization, with AI-ready labour in urban tech hubs coming out as winners and everyone else losing. Since human contributions to productivity will be relatively minimal, wages would likely flatten for most of those still employed.

Using software developers as an example, through the next five years, we’re likely to see the loss of between 40,000 and 160,000 jobs in the coding-adjacent technical workforce. That would represent between eight and 25 percent of current employment in the sector. That would likely also come with depressed wages for those left with jobs.

These are changes that are already in progress. I can’t imagine any scenario where I would ever need to hire a human programmer - unless he’s also willing to shovel my driveway and mow my lawn.¹

That’s one possible scenario. But it’s a pretty optimistic take.

The “Whoa! AI Sure Is Disruptive” Scenario

There are industry insiders who are convinced that things are moving a whole lot faster than most of us probably imagine. In this “worst-case” scenario, agentic AI tools are increasingly capable of independently executing complex tasks. This will allow significant labour substitution at the system level, replacing most human activity involving legal, technical, and financial operations.

I’m definitely seeing plenty of signs of this in my own work and I’ve heard about parallel developments across multiple sectors. Work that once required retaining a handful of lawyers for weeks or a team of analysts over a fiscal quarter can now be done in hours or days. At that point, demand for labour doesn’t decline. It collapses.

Now add in exponential growth in general-purpose robotics where technologies are increasingly commercially deployable for dangerous, dirty, and repetitive physical tasks. Think how many people are currently employed in warehouse, construction, mining, agriculture, and maintenance operations. Also: think about that from the perspective of a business owner whose major costs involve salaries, benefits, and healthcare expenses.

The sudden economic shock from all that would be unprecedented and devastating. We could see millions of jobs eliminated or radically deskilled. Sure, we’ll still need plenty of “human-in-the-loop” oversight jobs, but they’ll pay far less than the roles they’re replacing. Think how much less senior analysts now earning $120,000 will be able to demand once they’re renamed “AI supervisor”.

Lost income on that scale would lead to some significant system-wide pain, including:

• $50 billion less federal and provincial income tax revenue each year

• Maxed-out Employment Insurance systems

• 30 percent lower urban residential property values

• Spiking mortgage defaults

• Municipal council defaults

• <Your worst nightmare here>

How Might Canadian Governments Respond?

Based on decades of experience, you can expect governments to draw on many of the same go-to tools they used to manage previous crises:

• Skills retraining programs for the unemployed and underemployed. Such programs historically led to modest successes - even in economies where there was plenty of unmet demand for specific skills. That will likely be far less true this time around.

• Deep deficit spending to finance supplemental wage subsidies and EI expansion.

• Deeper deficit spending to subsidize critically damaged industries (which will often face no realistic hope of survival).

• New taxes, capital gains formulae, and usage levies on businesses driven by AI and robotic workflows. In principle, there’s certainly nothing wrong with taxing successful businesses to support social programs. But designing taxes that aren’t punitive to the point where they drive businesses out of the country can be complicated. And the performance of recent Canadian governments hasn’t been steller on that score.

• Rebalancing payroll tax systems. As with business taxes, this is a devilishly difficult needle to thread.

To be fair, there may not be a lot of great options available for even the smartest policy planners. But it’s a pretty good bet that what never worked well in the past will probably fare even worse this time.

What Should Governments Be Doing to Prepare?

While there doesn’t seem to have been a lot of thinking about these issues inside government, it’s not like no one has addressed the potential crisis. And, to be fair, at least one parliamentary committee has published thoughts on the subject. Although, hilariously, a primary focus of that discussion appears to have been assisting “diverse” workers. This is “arranging deck chairs on the Titanic” level prioritizing.

It would be nice to get some sense that the galaxy brains in Ottawa have at least dragged themselves out of bed to address the real issues.

At the same time, I would suggest that “positioning Canada” to “show leadership through global initiatives” is a distracting waste of time. The number of influential people and institutions on Earth who care what Canada thinks about these things is zero. And even if the Prime Minister did find some audiences keen on listening to his thoughts, the private businesses driving change are not among them.

Similarly, building AI compute infrastructure here in Canada in a way-past-its-prime attempt to create a local market is crazy talk. Elon Musk is just months away from deploying the first of tens of thousands of AI data centers into orbit where they’ll be powered by the sun. Do you really think a couple of overpriced warehouses somewhere in northern Quebec can compete? And that’s assuming such projects even received preliminary approval before 2038.

From everything I’ve seen, the number one priority right now should be eliminating government deficits and sharply reducing debt. Facing catastrophic tax revenue declines at the same time as catastrophic personal financial crises would be difficult even with room in the budget to maneuver. But heading into history’s greatest economic crisis with a $90 billion annual budget deficit is pretty much the same as vigorously waving a table cloth in surrender before the battle even begins.

---

1. As I wrote those words, it occurred to me that the charming gentleman I hired to rehabilitate my lawn last spring had, for many years, worked as a developer. Talk about proving my own point.