---
id: "concept-induced-demand"
type: "concept"
source_timestamps: ["¶4"]
tags: ["economics", "market-dynamics", "ai-economics"]
related: ["concept-complementarity", "claim-hinton-radiology-error", "entity-geoffrey-hinton", "prereq-microeconomics"]
definition: "The economic principle where a reduction in the cost or friction of a service leads to a disproportionate increase in the overall demand for that service."
sources: ["futures"]
sourceVaultSlug: "hbr-seg-futures"
originDay: 2
articleStem: "hbr-cl-84-big-tech-capability-crisis"
sourceUrl: "https://hbr.org/2026/06/big-techs-looming-capability-crisis"
sourceTitle: "Big Tech’s Looming Capability Crisis"
---
# Induced Demand in AI Economics

## Induced Demand in AI Economics

A fundamental economic principle: when the cost or friction of a service falls, demand for that service tends to rise disproportionately. The authors deploy this to explain why [[claim-hinton-radiology-error|Geoffrey Hinton's 2016 prediction]] that AI would replace radiologists failed.

Because AI made medical imaging **cheaper and faster**, the overall market for imaging *expanded* rather than contracting. Consequently each radiologist became more productive, and aggregate demand for their services **surged** rather than collapsed.

### The numbers (2025)
- Average radiologist pay: **$570,000** (up **9% year-over-year**)
- Persistent shortage: roughly **130 days to fill** a role
- Massive job-board demand

### Why it matters here
Induced demand is one of the two pillars (with [[concept-complementarity|complementarity]]) that the authors argue Big Tech is ignoring. If cheaper code expands the total surface of software that must be built, integrated, secured, and maintained, then demand for high-level [[concept-judgment-debt|engineering judgment]] rises rather than falls. See [[entity-geoffrey-hinton]] for the source of the failed analogy and [[prereq-microeconomics]] for the assumed background.

> Enrichment caveat: The radiology case is real, but note the counter-reading — induced demand does **not** guarantee employment growth in *every* AI-affected occupation, since software output is globally scalable and less regulated than medical diagnosis.


## Related across articles
- [[concept-ai-jevons-paradox]]
- [[claim-efficiency-increases-demand]]