Griffiths: Canada’s AI debate has a mile-wide blind spot. It’s our immigration policy

Good commentary and flagging a real issue that will become more important as AI develops. Just as our trade strategy has been slow to address IP and AI implications, so has immigration:

…But the labour market we are barreling toward is anything but normal.

If the AI thesis is even half-right, the bottleneck for Canadians over the next five years will not be a shortage of workers. It will be a shortage of jobs that AI cannot do—almost all of them physical, hands-on, or relational. Trades. Eldercare. Construction. Skilled installation. Personal services. The very segments where displaced white-collar Canadians, suddenly competing for “hands-on” work, will exert powerful downward pressure on wages. Adding hundreds of thousands of newcomers per year who will struggle to compete in either the contracting knowledge economy or in a skills economy experiencing a surfeit of labour, multiplies the AI disruption rather than relieving it.

The radical idea—and it is radical only because no one in Ottawa is willing to say it out loud—is that an AI-dominated economy may make Canada’s high-immigration model not just unnecessary but actively harmful to the workers we already have. The composition of intake, in such a world, would also need to flip towards ultra specialists, veritable immigrant unicorns, and away from generalist knowledge worker credentials in fields AI can now do at a hundredth of the cost.

Make no mistake: this argument is not nativism dressed in economic clothing. It is the opposite. It is the argument that immigration policy, like every other major lever of state, must respond to the actual economy in front of it—and the economy in front of us is being rewritten in real time by machines that, on a benchmark designed by the companies building them, now do the work of 44 professions about as well as the people who trained for 14 years to do it.

Canada’s policy elites have so far met this once-in-a-century technological inflection with something between a “let it rip” shrug and a quiet hope that the transition to an AI-dominated economy will be slow enough to manage the wrenching structural adjustment. It will not be. The AI dislocation looks set to be wider, deeper, and faster than the prepared playbook anticipates, and immigration policy will be the first of many non-labour files to get caught up in it.

Far better to have the difficult conversation now than after the displacement has begun in earnest. Time, as it has a habit of doing in moments like this, is fast running out.

Source: Canada’s AI debate has a mile-wide blind spot. It’s our immigration policy