TD Economics sees AI-induced 'job apocalypse' as unlikely despite concerns

TEXT START: TORONTO - A new report from TD Economics says fears of widespread job losses due to AI are unlikely, saying the labour market is more poised for gradual change instead.

The Dissection

The article does not disprove AI-driven displacement. It narrows “job apocalypse” to an immediate mass-unemployment event, then notes that current AI deployment has not yet reached the required threshold of autonomy, affordability, and rapid diffusion.

That is a near-term adoption assessment masquerading as a structural verdict. Its admission that transformation may proceed gradually and remain invisible for years undermines its own reassurance. Lag can conceal the approach of the break.

The Core Fallacy

The report mistakes slow arrival for non-arrival and current friction for permanent constraint. The Discontinuity Thesis does not require an overnight collapse. AI can first decompose jobs, compress headcount, weaken wages, and eliminate entry-level pathways while unemployment statistics remain deceptively calm.

“Reshaping work” is not benign if one worker supervises systems that replace five, then ten. Complementarity is a transitional phase, not proof of a durable human economic role.

The claim that productivity gains can facilitate job creation also imports a historical assumption that fails under cognitive automation dominance. Higher output can coexist with fewer workers. A 0.5–1 percentage-point productivity gain is not social rescue; it is evidence of greater displacement capacity unless ownership and income distribution change.

Hidden Assumptions

• AI must become fully autonomous before it can destroy jobs; partial automation across many tasks can hollow out entire roles.
• Adoption barriers are permanent constraints rather than temporary frictions intensified by competition and cost pressure.
• Tasks that are not financially feasible today will remain so as systems improve and workflows are redesigned.
• Jobs survive if some human tasks remain, ignoring headcount reduction and role recomposition.
• Aggregate job creation will absorb displaced workers at the necessary scale and speed.
• Canada’s low adoption is protection rather than lag, despite the article calling it an adoption gap.
• Unemployment is the decisive metric, excluding wage suppression, underemployment, degraded career ladders, and lost bargaining power.
• Gradual change is less dangerous. It may simply make the collapse harder to see until the labor market has already been reorganized.

Social Function

Partial truth functioning as ideological anesthetic and transition management. The report may accurately describe the present: adoption is uneven, autonomous execution remains limited, and an immediate employment cliff is not established by the supplied evidence. It then stretches those observations into reassurance about the trajectory.

Its institutional function is to defend the current “base case” and make structural rupture appear remote while firms continue testing the replacement mechanism.

The Verdict

This is not a refutation of the Discontinuity Thesis. It is a lag report. TD measures whether the guillotine has fallen; the thesis asks whether the blade is already being lowered.

The real threat is not one spectacular unemployment spike. It is the gradual severing of the mass employment → wage → consumption circuit through the erosion of human economic necessity. The article’s own evidence—slow diffusion, hidden transformation, concentrated exposure, productivity gains, and an adoption gap—describes the staging phase. “Gradual change” may simply be collapse spread thin enough to evade the obituary.