Businesses Are Using AI to Transform Work, Not Cut Jobs - Liberty Street Economics

TEXT START: The ongoing advancement and adoption of artificial intelligence continues to raise concerns about widespread job losses.

The Dissection

This is a six-month snapshot of regional firms, not a test of the final economic equilibrium. It records rapidly rising adoption—61 percent of service firms and 51 percent of manufacturers—but shallow investment, limited worker exposure, few immediate layoffs, reduced hiring at some firms, and widespread retraining.

The text then converts “no mass layoffs yet” into “AI is transforming work rather than eliminating it.” That is the central maneuver. It measures visible terminations while undercounting vacancy nonreplacement, natural attrition, shrinking entry-level intake, task compression, and higher output produced by fewer workers. The article itself admits that entry-level workers may already face significant barriers. That is not a side effect; it is an early signal of productive participation collapse.

The Core Fallacy

The article mistakes transitional lag for terminal stability.

AI augmentation and labor replacement are not opposites. Retraining an incumbent to automate routine work can be the mechanism by which future hiring becomes unnecessary. A firm can retain its current employees, increase their output, and still require fewer humans per unit of production. The protected incumbent workforce is not proof that the next workforce will be needed.

The survey also does not test the hardened framework. It does not establish whether AI has durable cost and performance superiority, whether human-only economic domains can survive competitive pressure, or whether the majority will retain access to economically necessary work. Three years of adoption data and six-month workforce windows cannot falsify the Discontinuity Thesis.

Hidden Assumptions

• Firms will preserve labor instead of using productivity gains to reduce costs, prices, hiring, or headcount intensity.
• Current employees and future entrants face the same labor-market conditions.
• Self-reported AI use captures all automated task substitution.
• Minimal investment today implies minimal future impact rather than early-stage diffusion.
• Retraining will permanently preserve roles instead of making fewer workers capable of doing more work.
• A human-in-the-loop requirement will remain economically substantial rather than becoming a thin supervisory layer.
• Low layoffs measure labor-market health, even though the more consequential variable may be whether new workers can enter at all.
• Results from New York and Northern New Jersey regional surveys can support a broader systemic conclusion.

Social Function

Partial truth functioning as ideological anesthetic and transition management.

The data may be accurate. The inference is structurally premature. The report reassures institutions by framing early displacement as workforce adaptation, elevates retraining as the solution, and avoids the ownership question: who controls the AI capital and receives the output when human labor is no longer broadly necessary? It is an institutional lullaby built from genuine statistics.

The Verdict

This article does not refute the Discontinuity Thesis. It documents an early diffusion phase in which incumbents are retained, firms experiment cheaply, and displacement appears first as weaker hiring and blocked entry rather than mass firing. The decisive question is not whether current employees survive the transition; it is whether firms need fewer humans to produce the same output and whether new cohorts can enter the system. On the supplied evidence, social death is not yet demonstrated—but the article provides no reason to downgrade the endpoint. It confuses the quiet phase of labor displacement with the absence of displacement.