AI creeps onto Fed radar, but footprint is small so far - The Globe and Mail

TEXT START: AI is storming through markets, construction and corporate planning.

The Dissection

The article is doing two things simultaneously: documenting the first measurable fractures from AI adoption and using their small statistical footprint to justify institutional delay. It treats AI as a future macroeconomic variable awaiting cleaner data rather than as a structural force already reorganizing ownership, labor demand, and bargaining power.

The evidence it cites is not trivial: AI is named in roughly 30–40% of layoffs where reasons are given, tech layoffs remain concentrated, insurance employment is falling, and AI-related hardware costs are spreading. But the article subordinates these signals to CPI weights, unemployment rates, and Fed meeting decisions. That converts a systemic transition into a measurement problem.

The Core Fallacy

The article assumes AI’s significance depends on whether it is large enough to move the Fed’s current policy dial. That is backwards. The Discontinuity Thesis is governed by the automation threshold, not by the visibility of the first aggregate statistic.

AI does not need to dominate CPI or unemployment before it becomes economically decisive. Firms can eliminate hiring, replace attrition, compress wages, and automate tasks long before official job-loss data records a mass liquidation. Falling unemployment can coexist with a collapsing future demand for human labor because labor statistics are lagging instruments built for a system that assumed human labor remained necessary.

The article also treats inflation and displacement as roughly symmetrical policy signals. They are not. AI capital spending can create immediate chip shortages and equipment inflation while simultaneously preparing the destruction of wage income. The temporary price pressure is visible; the consumption circuit it undermines is delayed. By the time the labor effect becomes undeniable, the ownership structure may already be locked in.

Hidden Assumptions

• Announced layoffs capture the main employment effect. They do not capture hiring freezes, nonreplacement, attrition, reduced hours, outsourcing, wage suppression, or the disappearance of jobs that were never created.
• A low unemployment rate proves labor-market strength rather than reflecting lagging data, sectoral churn, or temporary absorption.
• AI’s impact can be isolated cleanly as a single variable. In reality, its effects diffuse through productivity, pricing, capital allocation, staffing, and supply chains.
• The technology sector is merely an early victim rather than the advance unit of a broader cognitive automation campaign.
• Productivity gains will translate into broad household demand instead of concentrating income and control among AI-capital owners.
• The Fed can stabilize an economy after the wage-to-consumption circuit has been structurally severed.
• “Down the road” implies gradual adjustment rather than a threshold event after firms discover that human labor is no longer competitively required.
• Statistical invisibility means limited structural importance.

Social Function

Primary classification: transition management and ideological anesthetic, with a substantial partial truth.

The article is not fabricated. It accurately reports that the macro data has not yet cleanly registered AI’s full effects. Its anesthetic function lies in treating that absence of clean registration as grounds for patience. It reassures institutions that nothing decisive has happened because the dashboard has not yet turned red.

This is how the old order narrates its own approach to the cliff: the engine is failing, but the gauges remain within tolerance.

The Verdict

The article is a competent snapshot and a poor structural diagnosis. It correctly identifies early AI-driven layoffs, input inflation, and measurement failure, then mistakes the weakness of the signal for weakness of the force.

The Fed is not missing a minor policy input. It is watching a system that may eventually destroy the productive participation on which its models depend. When the aggregate data finally confirms the transition, the transition will already be advanced. The policy dial is quiet because the instruments are late—not because the machinery is healthy.